Pranjali: From Student to Startup Innovator

I don't talk to kids. I was just like, I have to do something. My parents had brought me here all the way from India and I had a little bit of that. I have to get into a good school. And also my parents had made it very clear that if I don't get a scholarship, if I don't have a scholarship then I won't be able to attend university over here. I would have to go back to India and manage. And that was because they both were students, PhD students at the time. So financially, for us as well, we were managing with roommates and stuff like that. At a young age, networking with top folks, taking them in as advisor, networking with Sam Altman before the AI hype. What do you think about networking? So hello everyone, today we have Pranjali here. She has been famous on all the Indian news media, titled as 100 crore startup creator in the US at the age of 16. So there's a lot to talk about not to... Breakdown, first of all, can you please introduce yourself and what did you what do you think about when you were highlighted on news media with this title? Of course. So nice to meet you all. I'm Pranjali. I'm currently building Delvehain. And yeah, I mean, I'm building this for past two and a half years now. The media that you're talking about, yeah, that happened last year. Yeah, I mean, that was quite interesting because I never imagined for this to spread in India as fast as it did. It started with... like a simple business insider article and we did it just because we were like basically preparing to raise money and media helps with that sort of stuff um and then one thing led to the other and then there were 600 million impressions in indian media wow we were just uh yeah we were just surprised by it so the source was business insider right you got featured on business insider and indian media picked on after that business insider yeah that was that was absolutely insane We're going to talk a lot about Delvei, how you built Delvei, how it reached this market cap and especially in your high school, when I was watching at the age of 16, maybe Shin Chan in my free time and with difficulty, assignments used to end and you thought of creating a startup. So before that, let's talk about your move from India to the US. So how was the move like, you know, moving in your middle school, I believe? Yeah, so I moved in seventh grade. My dad was in the army and we used to travel around a lot. My mom decided to come here for a PhD in public policy. And then we followed two years later. My dad ended up taking a sabbatical to also pursue his PhD here. And then we all moved to Miami. He's in Tech 2, right? So he was in the army and now he's doing his PhD in biomedical engineering. Pretty cool. So how did you get into coding? So I read that you started coding at the age of seven. Yeah. One business insider. Yeah, that was. So my. So it was, I mean, of course, my dad, he was. into tech and he was an engineer in the army so um a lot of news articles have extrapolated this as oh he was an engineer as well oh he owns a company there for sure yeah yeah so he's actually um just like he's basically just a nerd um and he got me into like scratch and basic coding stuff when i was young um and then i think that became more matured when i came to the us um and i went to um I mean I went to high school and middle school and he actually started doing competitive programming. In India? No here. Okay. So India it was mostly just you know things that my dad would introduce me to or you know just exploring things on my own on the internet just that and then coming here I found like more of a an environment that helped me hone those skills more. What can you describe the environment? Yeah so in middle school right I went to a middle school that was really like there was a community there there was math and tech focused um there were a bunch of indian kids um so yeah i mean we were teaching competitive math um you know compared the state and national level so that was there and then we also used to do programming competitions like programming team in middle school and then high school as well so i got reduced to like java python and more formalized programming that is incredible like because i think more schools are are famous for teaching programming at a young age in california and maybe in atlanta only the top school so did you go to like one of the top schools of florida it was a charter school actually so interesting story the first year that i moved here in seventh grade i changed three schools because uh the first two were just they were close to my i should know the first one was very it was local school um that environment wasn't great personally speaking it wasn't something that i like i think you should always surround yourself with people that are you You know, competing to be the best. And then you are also, you know, somehow have that motivation in you as well. So, that first school wasn't the right fit. Then my second school was an all-girls school. I had to travel like four hours. I mean, yeah, four hours every day combined. Two hours back and forth. Via bus? Via bus. In Miami? Yeah, in Miami. Because I... It was a very academic, only girls school. That ended up not being a great fit as well socially because I was, you know, just came from India. So it was really hard to accommodate. And yeah, that was not a good experience. And then finally moved to Doral Academy, which was a charter school. About 20-15 minutes from my house. And then I found this community of like math and science focused people. Incredible. So for all Indian immigrants watching, maybe coming with their kids to the US, can you describe what are charter schools? Yeah, of course. So charter schools are basically, I mean, my school was a lottery based admission system. So you just enter your name and, you know, maybe you will get in, maybe you won't. But most likely you will if you're living in an area close to it, about 20-25 minutes ish. And actually, no, the area, now that I think about it, area is not the biggest factor to get in. A charter school is just like a public school, but they have like a more organized system. They have... consulting companies are held like essentially public schools that are much better try to be better you know they have that incentive to keep improving their curriculum and you know and they're not as expensive as private schools you don't have to pay a huge fee so i see so like even in that kind of school they were you were encouraged to learn coding and i think you sharpened you mentioned in the article that you sharpened your skills at the age of 11 you were like be able to build stuff right can you describe at the age of 11 what what were you building The age of 11? I'm gonna think about it. I think 11, 12 may I was just doing like simple projects at home. I mean, I was learning programming. I would build like, you know, stuff that you see online all the time, you know, watch a tutorial to build something, build a calculator, build something, something fun. I was into animations and graphics. So, like, I would build like, like Star Trek, you know, graphics for those characters in the terminal and stuff like that. Yeah. So just playing around at the end of 11, I wasn't exactly building, building, building a full stack stuff. And then 12, 13 onwards, I started getting into more competitive programming. So it's like, you know, going to contests and then trying to solve, solve problems. DSL. Yeah, basically. And now to network, you need networking skills beyond, especially with mid mobile, because this mobile network in US gives you. everything unlimited starting at 15 per month and especially for travelers like me whether i go to india us canada mexico i have international roaming in my pocket at all times check out mid mobile by going to midmobile.com slash thing and get your deal immediately so at what point like you figured out you're special or maybe you did not like usually people describe confidences is having the faith that your ability is better than the standard yeah you That's a really interesting question. I've actually journaled about this a lot. I think having self-esteem and self-confidence, where does that come from? I think a lot of it maybe came from, I would love to thank my parents for this as well. I think as a kid, I was interested in a lot of different things. Every other month, I would have a new interest. Today, I want to do oil painting. Tomorrow, I want to do cooking. Next day, I want to do acting. And they would always encourage me for whatever it is that I'm interested in. it wasn't like oh you just have to study you just have to get good grades so having that sort of open mindset and encouraging me to do the best in whatever it is I'm doing I think that's what helped inculcate like that self-esteem and self-confidence initially can you break them down example with Anna like at what events that happened in your school that you were like yeah I am the best in in my field and I can do something beyond let's see I think uh the do something beyond and I'm the best feeling uh Yeah, that probably happened, let's see, I think that happened mostly like in when I came to the US in middle school and stuff, when I used to do math competitions. And I came from a completely different academic environment, right, from India. And I think in India, we're more encouraged to do rote learning and stuff like that. Here, initially, the first two schools I went into, I felt like the standard of education was much lower than what I thought. So it was very easy for you. It was very easy. But then as soon as I went to the charter school, there was there were a lot of competing Indians competing Basically, wherever there's a competition is you know a good standard So I was doing math and science competitions and the first few tests I took I did not do well because I think I was just adjusting to the American Contest system with the way that blue math math contest over here and then I just kept at it and eventually I ended up getting better than the rest of the people in You know who were there since like three two three years already so I think that's when I initially felt like okay I think if I have some sort of an innate level of skill combined with you know hard work or whatever can really take me to good places I think that's when it happened. Your roller coaster journey from the age of you know 11-12 to reaching that Miami Tech Conference where you presented your AI startup. So walk us through the rollercoaster journey from the age of 12 to 16. Of course. So 12, I'm doing all these math competitions. Slowly, I'm starting to get into hackathons due to the programming, co-better programming. So, you know, I used to do a lot of online hackathons, built a bunch of like very small projects in like AI and stuff, you know, mental health companion, this and that. And then eventually I discovered about Miami Hack Week. So my computer science teacher, actually, in eighth grade, or ninth grade, sorry. Yeah. He was telling me about Miami Hack Week, which actually one of his past students, his story is also insane, Michael Seaman. He's known as the app kid. Yeah. So when he was 14, his parents moved here from, I think it was Chile. To Miami. To Miami. To Miami. And, um... They were having a lot of financial troubles. So he took up on himself. He started making apps, pushing down the app store and started making money from them. And started to support his family at the age of 14, 15. Eventually, this app got so popular that Facebook ended up inviting him to their headquarters. And he started working for them. And he's worked for Facebook, Microsoft. No, not Microsoft. Facebook, Twitter, and Google, I believe. It's just 21, 22. Now it's started its own company as well. So no degree and this was his journey. Yeah, this was his journey. Was he your biggest inspiration? So at that point, he's still like one of my biggest inspirations. And yeah, I mean, he was going to come to that conference. So that's how I got interested and I ended up going to that conference. Exceptional because when you know if someone has done it, you feel like I can do it. You can do it as well, yeah. I was just like, I want to do something. Like my parents had brought me here from all the way from India. And... I had a little bit of that, you know, that I have to get into a good school. And also my parents had made it very clear that if I don't have a scholarship, then I won't be able to attend university over here. I would have to go back to India and manage. And that's because they both were students, PhD students at the time. So financially, for us as well, we were managing with roommates and stuff like that. So yeah, I had to get a scholarship. So there was a little bit of motivation from there that was kind of... Helping me, you know, or motivating me to look beyond just high school and look into other opportunities. And that's how I ended up finding out about these hackathons and going to them. So these hackathons were for teenagers? Like, anyone can go? Anyone can go. In Miami Hack Week specifically, I think there was an age requirement. There was an age requirement and I wasn't able to get in initially, actually, as a participant. I tried contacting so many people. did not end up getting it but then i just decided you can i found you can volunteer as well so i ended up going as a volunteer and helped them set up a lot of tents network with people and then um somehow found somebody that led me as a participant i remember when i was going to hackathons at the age of 18 19 to a lot of kids you know a lot of especially asian parents they bring the kids yeah to these hackathons and i never thought like they can they will be competing next to me so how do college students looked upon you as like you know maybe a teenager competing with them Actually, I think a lot of people were very encouraging actually they were like they were happy to see somebody young there you know presenting an idea at that sort of level. So during the hackathon before the actual hackathon started we were just pitching our ideas on the stage and trying to see if other people were interested in kind of getting a team and I remember at that point I didn't the startup idea wasn't there i was just there for the hackathon and i went and presented the idea and i was like i started with i'm in high school and that ended up getting the attention of a lot of people initially of course then the idea as well um i think i put something about put something under um i think it was a combination of climate change and ai at that point i just want to build something right and uh people who are interested will message me and um yeah At age 14, right? Age 14, yeah. And from here, how did you reach to a point when you had the expertise to build something with AI and something as good as Dell AI? And to teach people, first of all, what is Dell AI? Can you describe as well? The hackathon itself, I didn't have an idea at the beginning. But, you know, I was surrounded by not just college students, but YC founders, investors. And that was the thing that ended up getting me into startups. At hackathon? At the hackathon. Really? Yeah. Miami Hack Week. Wow. Yeah. So, it was a seven day long hackathon. And I think that's exactly what... It wasn't just a regular hackathon, it was my first in-person hackathon. which ended up being like VCs and almost everybody from the tech world. That's a real hackathon. Like at Georgia Tech, I have never seen like that many VCs coming. You know, this is like a seven day and you need more time for it. College students don't take seven day off. So what kind of students were there? So there were people from, in Miami, there were people from all over the country. There were, you know, college students from New York, from Stanford, from a lot of different universities. There were people who were already working at tech companies. um so it was it wasn't just a hackathon i guess i don't know how to describe it but um yeah it's like it's a perhaps a hackathon for you know tech professionals adults in a way you can call it that so from here you got to introduce how startups work how vcs work so you knew you can take it forward and build something and present to a vc as well yeah yeah i mean uh i didn't exactly learn that that's how we see that's how things work in startup world um i was just talking to a lot of people I ended up finding out that Lucy Guo, who was one of the judges there, she's also running... Co-founder of ScaleA. Yeah, yeah. She was, at the time, leading back-end capital. And they were doing a three-month residency in Miami, right after the hackathon. And they wanted... I think the top five teams were going to get interviewed to do HFZero. So I found out about that and I applied. I ended up interviewing for it. And I think at the last minute, initially they told me that there's no more spaces left. but I think I ended up getting in last mid somehow somebody dropped and I got in that was uh quite surprising for me but I think as a 14 year old it was quite the achievement and you met Sam Altman too right yes so how was that at what age you met I met him um what time 2022 August so after before chat GPT 3.5 after chat GPT 3 yeah beta yes yeah oh how was the how was the meeting like I mean, I met him at another conference. It was Future Forum. And I remember, I think, a lot of people were going up to him and asking him questions. And I remember just asking him, what do you think about the future of semantic search, you know, as a field? And he said that, yeah, like, that is one of those fields that you're going to see a lot of startups, a lot of growth coming in, in the future. And, yeah, I always remember that, you know, in the back of my head, that somebody that has created... Like, basically, the biggest foundation model that, you know, we're working with right now, somebody who's kind of leading AI as a whole, has that confidence in the thing that I'm working on. He was not a big face in AI at the time, right? He became the biggest face after December 2022. Yeah, I agree. That's because when Chachabee launched and stuff like that. But in SF and before that, he was obviously someone we already knew about, right? Like, open AI, you know. So, yeah, I mean, of course, now... The whole world is using Chan GPT, so there's an increased level of fame. But back then, he was a celebrity for me as well. This time also, you were living with your parents in a shared apartment because he was a PhD student, right? Because my parents were PhD students. Yeah, my dad was a PhD student. So how was that like for her? So that was in Miami, right? After HFZero, I ended up raising some money. So I would go to these different conferences, you know, San Francisco, Boston. I would crash at friends places and stuff. And that's how attendees conferences meet people. And your parents were always encouraging, right? Or you had to fight to get to convince them? No, definitely. I think once I kind of gave them the proof points, you know, that I've gotten somewhere, I have so and so funding, or somebody is vouching for the growth of whatever it is I'm doing. They were pretty, pretty encouraging. And in Business Insider, it was written that at the Tech Week Miami, like, You accidentally got introduced and funded. Was that true? Accidentally? I mean, not exactly accidentally. It was... I mean, I didn't plan to start a company, right? I was just there because Hackathon and meeting Michael Samen. So, at HF0, I went in with about 10, 11-ish ideas generally that I wanted. HF0 is BC, right? HF0 is the three-month residency that I did after the Hackathons. residency yes so it was um so back in capital was running hf0 it was three months founders living in one house and building a product yeah and they all get funded okay so in that in that incubator kind of yeah incubator so once you got the funding how much was the first funding that was actually hf0 now they're investing a million each per founder back then they were investing about 25k wow yeah wow we've grown a long way yeah you Okay, so from here you started living in an incubator. Yeah. Started living in an incubator. I ended up moving to virtual high school. So my entire education was now asynchronous and online. How did you manage? Like, you know, 11th, 12th, you talk to a JEE student, they cannot even finish assignments and you were building stuff. I mean, yeah, at that point, my parents have had this conversation with me multiple times, how different it would have been, 11th and 12th, personally, for me, if I was in India versus here. And yeah, I would probably be a JEE grinder if I wasn't here. But yeah, at that point, I think I consider... You know, the opportunities, I'm very grateful for the opportunities I got at that time. I think I 11th and 12th. Very interesting time. I think even compared to the average high school over here, right, they're all just like kind of on this track, getting into a good school. And I was just in a completely different space, you know, working, living with people who are 10 years older than me. So yeah, I think I was the second youngest person was 18 and then 20, 22 and so on. And you were 16, right? I was 14 in the incubator. In the incubator. Okay. So 14 to 16, you were in high school. Yeah. In incubator, managing high school. And at the time, you were not thinking about college, right? You were thinking about building and finishing assignments. Yeah. That's how you managed. Okay. So you've graduated high school at like, you're now 17. Yeah, now I'm 17. So you graduated high school. 18. Sorry, I just turned 18. Okay. So you graduated high school last year at the age of 16, 17. Yeah. So- Now, why college? Like, you said you were inspired by Michael Sammons, right? Who never went to college. Yeah, who never went to college. So, why college for you? I think somewhere along the line, I realized that being a solo founder is not great. You definitely need somebody to share that burden with. You need somebody to help you along the journey. And being in college, college is one of the best places to find a co-founder. You know, I used to read Paul Graham's essays. And I think it's something... said something like 40% of founders meet their co-founders in college. Yeah. And that is the biggest, you know, kind of, the place with the highest number of, how do I say this? Founder mode. Yeah, exactly. We find people here, especially at a college like Georgia Tech, I think the startup ecosystem is growing. And I'm finding, I'm meeting people who have actually done startups in high school as well over here. So a lot of builders, a lot of talented, you know, engineers. So I think it's one of the best places too. So before now, you were building everything alone? Yeah. Team of one? Really? I had engineers, of course, you know, on the team, employees and stuff, but I was the solo founder. So how big was your team when you were like in high school? About four people. So all were in India? Remote, remote. Some people were in Canada. One of them was in LA. One in India. Oh, and how did they feel about working for a teenager? I think they don't really care if they get paid. At the end of the day, that's really how I describe it. A lot of people ask me this. How did you work with people who are so much older than you? No one really cares if they get paid. So in your entire journey, it never happened that... I don't talk to kids That's a comment That's never happened That's never happened I think if you talk to like maybe old people right who are like kind of 60 plusish So you'll feel a little bit of a divide you know They feel like what is this kid going to do That type of thing they have that yeah but generally I've never felt that You said that you have lived in incubators but you still plan to go to college Isn't that a very expensive investment like 40k a year like especially because you're out of state right you came from florida and like you if you could find co-founders at an incubator wouldn't that be cheaper so uh i would have not gone to college if not for the scholarship i was i ended up getting a scholarship from master's own uh so that was the master's own foundation uh master yoshi son founder of softbank he has his own foundation Colmar Stone and it basically funds young people under the age of 28 and their education in the parts of their building. It's only for the US probably, right? No, it's international. Really? Yeah. So even if you're an international student coming from India, you could have been a legend. Exactly. That is what I... Like, I was an Indian citizen. There was no way for me to get these sorts of scholarships for US permanent residents or citizens or anything. So it was Coca-Cola's, you know, that... I think there's a scholarship for Coca-Cola. There's the Bill Gates Foundation. I was looking for all these options, nothing was there for me. So I ended up founding Masterzone. That is incredible. That's, I think, all international students need to hear. I definitely, I agree. Masterzone Foundation. How did you find? I, so through a friend, actually. So I used to go to, as I... told you I was going to San Francisco and Boston, these places. I ended up meeting somebody at Berkeley who is a student from Sri Lanka and he got this scholarship as well. Awesome. Now, in your journey, have you ever been like broke? Like, how did you fund the first biggest... Have I been broke? Yeah, in this journey. Like, how did you fix that? Like, you need capital to do at least something. For example, if you want to go to maybe San Francisco, you need $500 for the flight. I mean, I ended up raising money a little bit after. So, you know, HFZO happened and then I ended up getting some money from ODX on deck. And then AngelList as well, 2023, a couple more angel investors. So, that's what initially helped me out. So, once you got the first funding, you started fighting and then it happened. Awesome. So, now let's get to know more about your project. So, like for even for let's say a five-year-old kid can... can you walk them through Delve AI? Yeah, of course. So, I was actually just reading a book on this, Information and Society. And it was talking about how our entire world is, you know, information is growing at a rapid speed. And for a five-year-old kid, how would I explain this? Literally, for a five-year-old kid. Five to ten, you know. Like, think as smart as you, when you were ten. Think that way. I see. Because we want to make those ten-year-olds... old kids watching as well as smart as you are okay imagine this um you're really interested in a specific topic you want to know why the sky is blue um and you go to a library and you see 5 000 books there and you don't know which one to start reading to find the answer to your question how would you do that how would you approach that problem that's what dell solves so and you have a lot of competition in this space right you have Like similar projects could be consensus, right? That's a GPT, in chat GPT, like you can, you have questions about five research papers. Right. You can ask questions. Right. You have perplexity now. So how do you think you are unique? Right. I think search by itself right now looks very different than what it will look like in another 10 to 15 years. I'm looking long term, specifically just focusing on like the backend to improve search. I think both perplexity and consensus are built for like the broader consumer. you know there aren't a lot of tools that are built for enterprises or very specific use cases so we're focusing on consultants right example with anna specific use case um yeah of course i think consultants specifically um like they have like five or ten different uh data bases they have data in right they have like sales force have spot blah blah blah um then there's also public data they have to do research on like specific topic it's very disparate is what i'm trying to say um And that's why I made the analogy with the library, because they are that one specific customer segment for which information is so desperately stored in a lot of different places. For researchers, it is still consolidated in journals and research databases and stuff. But for consultants, it's a lot harder to pinpoint even what segment to look into to find the information that they're looking for. So that seems like a very complicated project. Like you have information coming from five different sources because Internet is organized thanks to. you know google search searching it but your personal company enterprise data which you mentioned is not organized right so how did you build up how did you build elf how did you solve this big problem right i mean we're still solving this problem right we are young startup which is 2.5 years old um but yeah at the moment uh initially two years like about for one and a half years of the initial uh journey uh we were building for researchers um and i quickly figured out that that problem is you know I think researchers as a market is a very tough market to crack and I'll admit this for anybody who's thinking about building for researchers they do not have the incentive to make their processes efficient internally because why think about it okay researchers is that one like researchers are that one specific industry where quality assurance is not a paid is not a paid position peer reviewers don't get paid to do the work that they do there's a lot of things that go wrong you can't treat researchers as like a regular industry that would invest in money to make things you know more productive why can you decide like yeah because maybe someone who hasn't done much research yes why they don't have peer review why they don't have qa yeah um they have peer reviewers but the peer reviewers are just just aren't paid it's not like that's just not how research research works um in research you don't pitch to get to get money you just write grant proposals and you have to stick to the grant proposal and you can't sway from it even if you think that there's something more novel out there. Like the way that research works is very different is what I'm trying to say. And for someone like Enterprise, I think it's a lot easier. to scale and build for them because they have like a very clear objective which is to just make the process more efficient reduce gray work and have as much output for the least input so research is a complex problem to solve because a lot of things are in development they're not proven yet and they are not because they can be proven because they are not used by hundreds of people so java ship jaldikar if things are shipped they are proven yeah but those things are not proven yet so that's why they are in gray area Yeah, I think that's one, like, I mean, obviously, research, it's R&D, it is still, there's a lot of things that are unknown, a lot of unknown variables there. So that's one thing. And of course, the way that the budgeting is done, you know, grant proposals, even the top research labs, they don't have a lot of just extra money around there, you know, to invest into other stuff. So I personally, like, this is what I thought that maybe I should start by focusing on just the top universities, and then I can create. products for like the broader consumer, broader market base. And I would just go to offices of Harvard Medical School professors or MIT professors and pitch it to them straight up. But I think people, I think finances were a big issue, even at those labs which have a lot of resources to invest into things that are outside of directly the scope of whatever it is that they're developing or investigating. So you found a big problem and now you're finding niche categories like university. Yeah. yeah okay so so how many customers you have so far um so we did about uh six enterprise pilots and currently we're about 250 plus individual users um and 250 plus 250 plus yeah and there's 50k 15k on the wait list that is still yet to have access to the product i have applied as well okay soon i will i will keep in the priority hi uh this is a demo to our current APIs, clustering, and key sections. And so I've run a query on COVID-19 sort of on the entire corpus of PubMed documents. And the source sort of serves as an indexer for all medical-related documents. And three or four customers are actually using it at the moment. And so I've asked a question around autoimmune response to COVID-19 patients when they specifically have Down syndrome. And so to index the documents that specifically talk about that relationship, while also understanding all the underlying topics and then doing second degree analysis, that's sort of what I'm going to show you. And so it's come up with around 6,000 documents, just under 6,000 documents for this particular query, close to 200 themes and topics and sort of clustered all these documents out in these little rectangles, bigger sections, bigger rectangles and going smaller depending on. how many times that particular theme occurs in that entire corpus. And so I have all the documents in the scope on the right. Let's say I want to understand maybe thrombosis or biomarkers or whatever better. Under that particular query, I can click Analyze Second Time, and then it runs the same theme and topic analysis on that particular query under the 6,000 documents. And so you have a new tree map of all those topics. under that one sub-topic will look like. And so yeah, you have a second degree analysis and that's what the clustering looks like. The second part is key sections, which is once you have a document that you wanna go deep under. So what I previously showed you is like a lateral view of the entire domain. Now, once you have one document that you think is like very apt or you want to ask for, a specific question to, you just upload it here or even just give sort of like an open source link and then you can do Q&A on it. You can do one question on multiple files so you sort of save time when let's say for example somebody's doing a literature review, so they get all the paragraphs that are semantically associated with that query that directly answer that question in all of the documents that you uploaded. And so this is my account on the platforms I can show you previous queries that I've run. So it sort of saves everything in stored results. And yeah, I can sort of just go here and see what that looks like when you run a particular query on a doc. And so, yeah, so it's extracted those paragraphs that answer that question directly. My question was how good are neural networks at writing legal text? And that's the PDF that I wanted to ask from and sort of created. a coagulation of all the paragraphs that directly answer the question. And so, yeah, that's what the product looks like. You push your product to product hunt as well. So I think you were actively looking at product hunt. Did you build this to go after the hype or you were genuinely interested? I think... Or both. So you're mentioning AI specifically? Yeah, specifically because everyone is building stuff since 2022. So I started about the end of 2021 and everybody in my cohort out of... 11 companies right, the other 10 companies they were all crypto. I was the only AI company back then. So I like to think that I was there before the hype cycle came on. Yeah, so I think that's definitely there. And it was a core problem because I used to do research in high school and then that's where the problem kind of stemmed from. That's a great way to look at it. Like, you know, I think the Perplexity CEO just posted as well. Like, what is the top product of that year and what was the hype? VCs hyped crypto, but the top product was something else. You know, that shows that if you just go after the hype, maybe you might not be able to make it. But your hype may come later. Right, right, right. I definitely agree with that. And mine came very soon, right after I launched. But yeah, at the moment, I wasn't really seeing a lot of AI companies. Product Hunt was like a lot of crypto companies, you know, increase your savings by, you know, a bigger APR. by investing it in crypto and stuff like that. So a lot of ideas in that field, decentralized finance. And then AI got on pretty soon after that. And what were the biggest lows and highs of your journey? I think, let's see. I think biggest highs, there were highs and lows and then highs again and lows again. Highs definitely, you know, initially getting that, being in HL0, pushing to product hunt. And I will say this about HF0, it was probably a very productive time. As for an incubator, for any other hacker house that I've lived in, I've never, you know, kind of felt this sort of like infectious energy going on, you know, just like putting hours and hours into the project, bouncing ideas off of each other and just working hard. I would see people working until 6am every morning in HF0. So that would, as a 14 year old, you know, seeing that sort of environment, it pushed me as well to do the best I could. There was no excuse because I was the youngest or whatever. I had to do good. So, yeah, I think that was one of the moments I did a lot of work, product hand happened, ODX happened, getting investment from them. And then, of course, the tough times came on. I had, you know, an engineer. We were working on a product. We tried to raise funding towards the end of 2022, but then there was a little bit of a recession. So it was harder to raise money. So we weren't able to close that round. That was. kind of a low point I would imagine. And then, 2023-ish, my first hire ended up not working out, we had to let go. But we ended up continuing the cycle of growth, that was something that I kept consistently working on. And then, of course, the media impressions happened, a lot of people came on board, we got a lot of attention towards our product and now we're consistently growing. and I have a team of overseas engineers that are helping me work on the product. So it is sustainably growing. Biggest low, tell us. Maybe you thought that you took the wrong decision. Maybe the startup is going nowhere. Let's see. Yeah, I think we were going to close an enterprise contract in early 2023. University or company? It was a university, I remember. And at the same time, I was about to... I was pitching to VCs for a seed round as well. That contract ended up falling through at the very last moment and so did the VC funding. And that was one of the lowest points, I imagine, of my entire time building this company. And three, four months of just, like, you know, kind of not even having the motivation to, like, get started with the day. A lot of, it was quite difficult, I think back to that time. But, I mean, of course, my parents, they were my biggest support system. And I made a lot of friends through the teenage startup community as well, through different events. So, you know, that helped me push through. So, let's say if you have to look back, you're giving advice to a 14-year-old Pranjali, what would you say? I think I would just say, do what excites and don't be afraid. of, you know, what may happen. Just keep doing whatever it is that, you know, excites you the most, that you think. Yeah. I think that's literally in three words, do what excites. Amazing. Let's come back to the Dell VI, how internally, how the tech works. So can you break down the engineering? It has RAG systems. At what time do you use AI? So we have, we're using proprietary models, you know, fine tuning Lama and Mistral and using that to do RAG on very specific use cases. So that's the AI part. And then, of course, as the user uses it more, they're able to read the responses and we use that feedback to improve the model over time. Lama and Mistral. So why Lama and Mistral? How did you choose those? Yeah, I think because Lama and Mistral, of course, you know, launched by Facebook, open source model. I tweaked around with it, ran a few fine-tuning experiments just on my own, on like biomedical Q&A, biomedical question and answering tests. And it was doing really well. And I was just surprised that Lama has that sort of ability to, you know, I gave it about 100 or 200 odd examples and it was able to generalize and answer questions. So, you know, I think that's another thing that I would have good advice to young founders who are building in AI. Then build your own conviction on things, read papers and actually do the background research whenever you're deciding to, you know, take a step in your product or whatever. Do your own tests. For example, there is a very popular tool called HeyGen. I was making videos, just trying if HeyGen is for me or not. You give a voiceover, it does the lip sync. And in my same clothes, which I've trained them all before, it does it perfectly. But then I realized it's not fun anymore. So I think people have to go and try whatever the cool tools are there. Just don't listen to someone's feedback. For sure. Build your own conviction. Absolutely. So you got the conviction and what else is going in the tool? Yeah. So, RAG, of course, we have RAG there as well. I've been experimenting with hierarchical RAG. That's what it's called. So, I used to use Pinecone initially. I ended up looking at the documentation of Pinecone. Sorry. Pinecone is a vector database. So, it's used for semantic search. You know, store data, textual data. Like vector database. Yeah, basically. And they have a system called hierarchical embeddings, where basically you use... The concept of knowledge graphs to make search easier and more efficient. So imagine this, instead of just having your vector data embeddings just disparately stored, you cluster them first, like you create sub-communities and you create communities on top of them and like a kind of hierarchical system essentially. So we use that to sort of make a lot of data. Data essentially is dense. when you're thinking about doing research or consulting or whatever it is so um higher he collected database uh for rag um proprietary models and then of course encryption and templating to protect um you know customer data impressive and overall how long if a student watching this wants to build this how long it takes and if you have to rebuild it how would it how much it would take oh If I was to rebuild it, I would definitely not take as much time because now I have all the experience, you know, like, at least the knowledge that I didn't have before. A lot of trial and error happened, you know, before I figured out that these were the tools to use to build something. And the thing is that I think when you're solving a complicated problem, they always say just build an MVP, but build a high quality MVP. Don't build just like a regular, you know, like... hardcore yeah essentially um like i i used to use a cosine similarity i think um initially and ended up realizing that that would go out very fast um cosine similarity is essentially another technique in virtual databases to find similar like similar text to do search and um yeah it's an it's an easy way to do search but um you always have to think about what's the best what's the best uh thing to do for your user how can you deliver value to them in the most like less time and for that you may have to do more research and not just take you know the first solution that shows up which in my case was initially cosine similarity and then i did research and found out about hierarchical embeddings and all of that i see so trying different solutions is very important so figured out that cosine similarity is not the best way hierarchical one hierarchical embeddings i think it's just doing a lot of research initially i didn't really look into the back end of machine learning and do all that study work because you know i was just focused on building talking to customers and just doing the tasks and i think taking out time in like recently i've been taking out more time to actually look into the back end of uh of the product do research on by myself on machine learning and um yeah I think that's helping me definitely. I see. An interesting question could be, what are the technical expertise you have versus don't have? Like, what are the coding knowledge you have in AI, machine learning, or maybe DSA, and what you don't have? I think it's easy. It's important to accept what you know versus not know. Yeah, yeah, 100%. So can you break it down? What do you know and what you don't know? Right now? Yeah. And what are you learning, you know? Yeah. I think over the years of building Delve, I've learned a lot about full stack. And now I'm definitely learning more about machine learning. And as for what I don't know, you never completely know what you don't know. That's always there. But yeah, I think for now, I'm going to continue on this exploratory journey of NLP and ML and then figure out what next to learn from there. So what do you think is the best way for kids to learn programming, machine learning, AI? Yeah. I mean for me it was getting a little bit of head start initially with you know comparative programming and DSA and all that and then eventually getting into building my own projects. For young kids I would honestly I would encourage them to look into building their own projects and then first of all starting with like a very basic language, building something simple and then think about what would they would want to add on top of it and then learning new languages to do that. So it's like utility based programming like um what's what's the word? Inclement? Bottom up. Bottom up? Yeah. Example? Example? Uh let's see. Mujhe game banane hai. Uh I create like the most like basic version of it then I decide I want to add something on top of it and then eventually I just keep learning things to make whatever whatever it is that I want. I think that's the that's one of the best ways to learn stuff I think. I see do what you enjoy and build on top of it like slowly add more features make it more complex because it serves as motivation to like you know keep going um i think a lot of people learn for the sake of getting a job but you know eventually that's what ends up happening but if you learn uh if you keep what you enjoy it you know at the front of the forefront then you know you don't need external motivation to push you now coming back to college degree which you first thought you will not pursue inspired by yeah what are the things you're liking about your degree georgia tech experience so far and do you still think it's worth it um are you able to find your co-founder yeah it's only been the first semester to start to be exchanged i definitely met a lot of interesting people you know you um chris klaus and i've gotten like access to the startup community over here so yeah i think that's been a big support currently i'm still building on my own i am going to look out for my co-founder i'm just building actually I have met a lot of cool people here and been going to hackathons with them. So I'm sure, you know, soon, soon I will find someone who's the right fit. I remember Startup Exchange Competition at Georgia Tech. There were three winners. All three were Indian. And they're also Indian citizens, not even Indian Americans. Yeah. All three were Indian citizens. Yeah. So can you share, like, what is something in us now that we are on the forefront of everything? Like, we are, for example... whether it's like top CEOs of Silicon Valley. Yeah. We are there. Top, top, you know, entrepreneurship competition in the US. We are there. Maybe it's a small sample size I have, but what is in us? That's a good question. I think we have, I think we have a lot of fire inside us. We have a lot of ambition generally. I think. I mean there's a little bit of a sad duality to this as well. I think because we come from, a lot of people come from middle class backgrounds in India, I did as well. So sometimes it is about proving something to someone and it's not necessarily internally there. But I like to believe that us as a country, we have a lot of fire in us and we have the ability to work hard. That's why we're also, you know, the... I'm pretty sure the highest earning demographic in the US as well. Like you know post 120k, next is straight 90. Yeah exactly, look at that. Yeah so on top of that I would like to share that I think we are really good at finding out the hype, things that work like for example a lot of students are maybe not going to get jobs as easily with a college degree they used to. So now the next good thing you can do is to catch the hype of building stuff and they are good and in finding opportunities wherever they are possible yeah and that's how you see so many founders yeah i agree about being resourceful I think that is the key. That's the key. It's not what you have, it's what you do with what you have. And I think that's what my parents also instilled in me, I think, as a kid. Being resourceful is one of the biggest tools that you have over here. From finding hackathons to go to, to finding scholarships, to basically figuring out where it is that I can get what I want. Who are the most inspiring founders? We think a lot about the CEO of AMD, Lisa Su. She's definitely one of those very, very inspirational people now that I think about it. Initially, you know, as a young kid, I never really thought about, you know, sort of the gender aspect of building in tech. Now that I'm getting older, I'm in college, I see that more clearly. And this is why I respect her a lot. Of course, Sam Altman, for what he's built, he's leading, you know, the change in Gen AI. Who else? who is the ex-CEO of GitHub, Nat Friedman. Nat Friedman is definitely somebody that I look up to. Yeah, these are three names that come to my mind. Why? Can you share the reason too? Yeah, of course. I used to look up to his knowledge. Oh, it's Patrick Collison, CEO of Stripe. And Nat Friedman and Patrick Collison, both of them specifically because I spend a lot of time on their websites. I read a lot about, you know, advice, things to do before you're 20 on Patrick Olson's website, on Nat Friedman's website. And then he has, if you notice something, a lot of people in tech have very simple websites, you know, where they have their insights and all that. Yeah, like Patrick. And yeah, I think just having that distilled, just the knowledge that they've gained throughout the years, having access to that is something that I really, I really, you know, read them, read it a lot. Yeah. And, you know, you got... The top advisors on your board, for example, Lucy Gao, who founded of Scale.ai, who got kicked out. That's a sad story. So I think, what do you think about networking? At a young age, networking with top folks, taking them as advisors, networking with Sam Altman, before the AI hype. What do you think about networking? What do I think about networking? How do you partner with these people? Every single place where I've been able to meet these people, anybody could have gone. That is something interesting to pick. Future Forum, the conference that I ended up meeting Sam Alpinen or Miami Hackey, we were going to meet Lucy. Anybody could have gotten in. But it's about taking the initiative to find those things and go there. And I think sometimes it's not just about meeting somebody big. It's about finding whatever it is that you enjoy and finding those niche players that you really want to connect to. For me, they were these people. But yeah, I never planned to meet Sam Altman, yeah, planned to meet Lucy Goh. I just ended up being there because I was interested in the topics and then that's how it happened. And then you showed interest to them and they cared about you. Yeah, I think Lucy Goh, that definitely happened because of, you know, HFZero. I was building stuff, I had some good ideas and, you know, that didn't, however that happened, she saw value in what we were building, so yeah, that's how that happened. And for Sam Altman, it was just... asking him a random question and wasn't necessarily pitching my company to him um but even that that interaction to date you know it's quite special to me the big question a lot of founders are on o1 visa but you are a student are you on f1 visa which visa are you and how you got into it yeah um so i'm going to an o1 visa um initially i was dependent of my parents my um my parents were here for the for the phd so i was not dependent of an you know for student which was f2 then moved to h4 which is dependent of an h1b And then through Delve, I applied for O1. And that's how. How long the process took? So I took a break in the middle of the process while the legal stuff was going on because I was more invested in like, you know, like deals and was just really busy in general. So I put a pause on it. But if you just do it in one go, it should not take more than like two, two and a half, three months. Really? That fast? It is pretty fast. It has a two week turnaround from sending the application to when you hear back from them. That's amazing. You apply for Express? I apply for Lexstra, yeah. Awesome. And like on O1 now, you have freedom to do internships if you want to. Yeah, yeah. So whatever you want to do in your college journey as a founder, you have freedom to. It's basically an employment visa. Yeah. Like any advice you have for O1? Yeah, so there's some companies that are new startups actually that are serving founders that want to do O1. One of them is, I don't remember the full name, but it started Plymouth. And... I use LegalPad. It's a startup that ended up getting acquired by Deal, another company. But yeah, LegalPad helped me. And the average cost comes out around 2K, 10K-ish. 10K-ish, yes. Yeah, yeah, yeah. For O1. That's pretty good because like your entire immigration journey is straightforward after that. And even as an Indian, after O1, you get green card like within few years. Yeah, yeah, yeah. So that's amazing. Yeah, yeah. Actually, there's one, someone I know, Sigil Nguyen. He's another young... he's pretty famous in the young founder community he has this community and then he gives out hoodies to people who are recent recipients of the O1 it's called the Extraordinary Club I think something something like that but yeah there's a lot of resources on his website I can link share you sure thank you so much for sharing everything in detail and you know inspiring all the kids now to build at a young age and I hope people take away how to go to college fully funded and how to build stuff thank you so much again of course

I would have to go back to India and manage. And that was because they both were students, PhD students at the time. So financially, for us as well, we were managing with roommates and stuff like that. At a young age, networking with top folks, taking them in as advisor, networking with Sam Altman before the AI hype.

What do you think about networking? So hello everyone, today we have Pranjali here. She has been famous on all the Indian news media, titled as 100 crore startup creator in the US at the age of 16. So there's a lot to talk about not to...

Breakdown, first of all, can you please introduce yourself and what did you what do you think about when you were highlighted on news media with this title? Of course. So nice to meet you all. I'm Pranjali. I'm currently building Delvehain.

And yeah, I mean, I'm building this for past two and a half years now. The media that you're talking about, yeah, that happened last year. Yeah, I mean, that was quite interesting because I never imagined for this to spread in India as fast as it did.

It started with... like a simple business insider article and we did it just because we were like basically preparing to raise money and media helps with that sort of stuff um and then one thing led to the other and then there were 600 million impressions in indian media wow we were just uh yeah we were just surprised by it so the source was business insider right you got featured on business insider and indian media picked on after that business insider yeah that was that was absolutely insane We're going to talk a lot about Delvei, how you built Delvei, how it reached this market cap and especially in your high school, when I was watching at the age of 16, maybe Shin Chan in my free time and with difficulty, assignments used to end and you thought of creating a startup. So before that, let's talk about your move from India to the US. So how was the move like, you know, moving in your middle school, I believe? Yeah, so I moved in seventh grade.

My dad was in the army and we used to travel around a lot. My mom decided to come here for a PhD in public policy. And then we followed two years later. My dad ended up taking a sabbatical to also pursue his PhD here. And then we all moved to Miami.

He's in Tech 2, right? So he was in the army and now he's doing his PhD in biomedical engineering. Pretty cool.

So how did you get into coding? So I read that you started coding at the age of seven. Yeah.

One business insider. Yeah, that was. So my.

So it was, I mean, of course, my dad, he was. into tech and he was an engineer in the army so um a lot of news articles have extrapolated this as oh he was an engineer as well oh he owns a company there for sure yeah yeah so he's actually um just like he's basically just a nerd um and he got me into like scratch and basic coding stuff when i was young um and then i think that became more matured when i came to the us um and i went to um I mean I went to high school and middle school and he actually started doing competitive programming. In India?

No here. Okay. So India it was mostly just you know things that my dad would introduce me to or you know just exploring things on my own on the internet just that and then coming here I found like more of a an environment that helped me hone those skills more.

What can you describe the environment? Yeah so in middle school right I went to a middle school that was really like there was a community there there was math and tech focused um there were a bunch of indian kids um so yeah i mean we were teaching competitive math um you know compared the state and national level so that was there and then we also used to do programming competitions like programming team in middle school and then high school as well so i got reduced to like java python and more formalized programming that is incredible like because i think more schools are are famous for teaching programming at a young age in california and maybe in atlanta only the top school so did you go to like one of the top schools of florida it was a charter school actually so interesting story the first year that i moved here in seventh grade i changed three schools because uh the first two were just they were close to my i should know the first one was very it was local school um that environment wasn't great personally speaking it wasn't something that i like i think you should always surround yourself with people that are you You know, competing to be the best. And then you are also, you know, somehow have that motivation in you as well.

So, that first school wasn't the right fit. Then my second school was an all-girls school. I had to travel like four hours. I mean, yeah, four hours every day combined. Two hours back and forth.

Via bus? Via bus. In Miami? Yeah, in Miami. Because I... It was a very academic, only girls school.

That ended up not being a great fit as well socially because I was, you know, just came from India. So it was really hard to accommodate. And yeah, that was not a good experience.

And then finally moved to Doral Academy, which was a charter school. About 20-15 minutes from my house. And then I found this community of like math and science focused people. Incredible.

So for all Indian immigrants watching, maybe coming with their kids to the US, can you describe what are charter schools? Yeah, of course. So charter schools are basically, I mean, my school was a lottery based admission system. So you just enter your name and, you know, maybe you will get in, maybe you won't.

But most likely you will if you're living in an area close to it, about 20-25 minutes ish. And actually, no, the area, now that I think about it, area is not the biggest factor to get in. A charter school is just like a public school, but they have like a more organized system. They have... consulting companies are held like essentially public schools that are much better try to be better you know they have that incentive to keep improving their curriculum and you know and they're not as expensive as private schools you don't have to pay a huge fee so i see so like even in that kind of school they were you were encouraged to learn coding and i think you sharpened you mentioned in the article that you sharpened your skills at the age of 11 you were like be able to build stuff right can you describe at the age of 11 what what were you building The age of 11?

I'm gonna think about it. I think 11, 12 may I was just doing like simple projects at home. I mean, I was learning programming. I would build like, you know, stuff that you see online all the time, you know, watch a tutorial to build something, build a calculator, build something, something fun. I was into animations and graphics.

So, like, I would build like, like Star Trek, you know, graphics for those characters in the terminal and stuff like that. Yeah. So just playing around at the end of 11, I wasn't exactly building, building, building a full stack stuff. And then 12, 13 onwards, I started getting into more competitive programming.

So it's like, you know, going to contests and then trying to solve, solve problems. DSL. Yeah, basically. And now to network, you need networking skills beyond, especially with mid mobile, because this mobile network in US gives you. everything unlimited starting at 15 per month and especially for travelers like me whether i go to india us canada mexico i have international roaming in my pocket at all times check out mid mobile by going to midmobile.com slash thing and get your deal immediately so at what point like you figured out you're special or maybe you did not like usually people describe confidences is having the faith that your ability is better than the standard yeah you That's a really interesting question.

I've actually journaled about this a lot. I think having self-esteem and self-confidence, where does that come from? I think a lot of it maybe came from, I would love to thank my parents for this as well.

I think as a kid, I was interested in a lot of different things. Every other month, I would have a new interest. Today, I want to do oil painting.

Tomorrow, I want to do cooking. Next day, I want to do acting. And they would always encourage me for whatever it is that I'm interested in. it wasn't like oh you just have to study you just have to get good grades so having that sort of open mindset and encouraging me to do the best in whatever it is I'm doing I think that's what helped inculcate like that self-esteem and self-confidence initially can you break them down example with Anna like at what events that happened in your school that you were like yeah I am the best in in my field and I can do something beyond let's see I think uh the do something beyond and I'm the best feeling uh Yeah, that probably happened, let's see, I think that happened mostly like in when I came to the US in middle school and stuff, when I used to do math competitions.

And I came from a completely different academic environment, right, from India. And I think in India, we're more encouraged to do rote learning and stuff like that. Here, initially, the first two schools I went into, I felt like the standard of education was much lower than what I thought.

So it was very easy for you. It was very easy. But then as soon as I went to the charter school, there was there were a lot of competing Indians competing Basically, wherever there's a competition is you know a good standard So I was doing math and science competitions and the first few tests I took I did not do well because I think I was just adjusting to the American Contest system with the way that blue math math contest over here and then I just kept at it and eventually I ended up getting better than the rest of the people in You know who were there since like three two three years already so I think that's when I initially felt like okay I think if I have some sort of an innate level of skill combined with you know hard work or whatever can really take me to good places I think that's when it happened. Your roller coaster journey from the age of you know 11-12 to reaching that Miami Tech Conference where you presented your AI startup. So walk us through the rollercoaster journey from the age of 12 to 16. Of course.

So 12, I'm doing all these math competitions. Slowly, I'm starting to get into hackathons due to the programming, co-better programming. So, you know, I used to do a lot of online hackathons, built a bunch of like very small projects in like AI and stuff, you know, mental health companion, this and that. And then eventually I discovered about Miami Hack Week.

So my computer science teacher, actually, in eighth grade, or ninth grade, sorry. Yeah. He was telling me about Miami Hack Week, which actually one of his past students, his story is also insane, Michael Seaman.

He's known as the app kid. Yeah. So when he was 14, his parents moved here from, I think it was Chile.

To Miami. To Miami. To Miami. And, um...

They were having a lot of financial troubles. So he took up on himself. He started making apps, pushing down the app store and started making money from them.

And started to support his family at the age of 14, 15. Eventually, this app got so popular that Facebook ended up inviting him to their headquarters. And he started working for them. And he's worked for Facebook, Microsoft.

No, not Microsoft. Facebook, Twitter, and Google, I believe. It's just 21, 22. Now it's started its own company as well.

So no degree and this was his journey. Yeah, this was his journey. Was he your biggest inspiration? So at that point, he's still like one of my biggest inspirations.

And yeah, I mean, he was going to come to that conference. So that's how I got interested and I ended up going to that conference. Exceptional because when you know if someone has done it, you feel like I can do it. You can do it as well, yeah.

I was just like, I want to do something. Like my parents had brought me here from all the way from India. And...

I had a little bit of that, you know, that I have to get into a good school. And also my parents had made it very clear that if I don't have a scholarship, then I won't be able to attend university over here. I would have to go back to India and manage. And that's because they both were students, PhD students at the time. So financially, for us as well, we were managing with roommates and stuff like that.

So yeah, I had to get a scholarship. So there was a little bit of motivation from there that was kind of... Helping me, you know, or motivating me to look beyond just high school and look into other opportunities. And that's how I ended up finding out about these hackathons and going to them. So these hackathons were for teenagers?

Like, anyone can go? Anyone can go. In Miami Hack Week specifically, I think there was an age requirement.

There was an age requirement and I wasn't able to get in initially, actually, as a participant. I tried contacting so many people. did not end up getting it but then i just decided you can i found you can volunteer as well so i ended up going as a volunteer and helped them set up a lot of tents network with people and then um somehow found somebody that led me as a participant i remember when i was going to hackathons at the age of 18 19 to a lot of kids you know a lot of especially asian parents they bring the kids yeah to these hackathons and i never thought like they can they will be competing next to me so how do college students looked upon you as like you know maybe a teenager competing with them Actually, I think a lot of people were very encouraging actually they were like they were happy to see somebody young there you know presenting an idea at that sort of level.

So during the hackathon before the actual hackathon started we were just pitching our ideas on the stage and trying to see if other people were interested in kind of getting a team and I remember at that point I didn't the startup idea wasn't there i was just there for the hackathon and i went and presented the idea and i was like i started with i'm in high school and that ended up getting the attention of a lot of people initially of course then the idea as well um i think i put something about put something under um i think it was a combination of climate change and ai at that point i just want to build something right and uh people who are interested will message me and um yeah At age 14, right? Age 14, yeah. And from here, how did you reach to a point when you had the expertise to build something with AI and something as good as Dell AI?

And to teach people, first of all, what is Dell AI? Can you describe as well? The hackathon itself, I didn't have an idea at the beginning. But, you know, I was surrounded by not just college students, but YC founders, investors. And that was the thing that ended up getting me into startups.

At hackathon? At the hackathon. Really?

Yeah. Miami Hack Week. Wow.

Yeah. So, it was a seven day long hackathon. And I think that's exactly what...

It wasn't just a regular hackathon, it was my first in-person hackathon. which ended up being like VCs and almost everybody from the tech world. That's a real hackathon.

Like at Georgia Tech, I have never seen like that many VCs coming. You know, this is like a seven day and you need more time for it. College students don't take seven day off.

So what kind of students were there? So there were people from, in Miami, there were people from all over the country. There were, you know, college students from New York, from Stanford, from a lot of different universities.

There were people who were already working at tech companies. um so it was it wasn't just a hackathon i guess i don't know how to describe it but um yeah it's like it's a perhaps a hackathon for you know tech professionals adults in a way you can call it that so from here you got to introduce how startups work how vcs work so you knew you can take it forward and build something and present to a vc as well yeah yeah i mean uh i didn't exactly learn that that's how we see that's how things work in startup world um i was just talking to a lot of people I ended up finding out that Lucy Guo, who was one of the judges there, she's also running... Co-founder of ScaleA. Yeah, yeah.

She was, at the time, leading back-end capital. And they were doing a three-month residency in Miami, right after the hackathon. And they wanted... I think the top five teams were going to get interviewed to do HFZero.

So I found out about that and I applied. I ended up interviewing for it. And I think at the last minute, initially they told me that there's no more spaces left.

but I think I ended up getting in last mid somehow somebody dropped and I got in that was uh quite surprising for me but I think as a 14 year old it was quite the achievement and you met Sam Altman too right yes so how was that at what age you met I met him um what time 2022 August so after before chat GPT 3.5 after chat GPT 3 yeah beta yes yeah oh how was the how was the meeting like I mean, I met him at another conference. It was Future Forum. And I remember, I think, a lot of people were going up to him and asking him questions.

And I remember just asking him, what do you think about the future of semantic search, you know, as a field? And he said that, yeah, like, that is one of those fields that you're going to see a lot of startups, a lot of growth coming in, in the future. And, yeah, I always remember that, you know, in the back of my head, that somebody that has created... Like, basically, the biggest foundation model that, you know, we're working with right now, somebody who's kind of leading AI as a whole, has that confidence in the thing that I'm working on.

He was not a big face in AI at the time, right? He became the biggest face after December 2022. Yeah, I agree. That's because when Chachabee launched and stuff like that. But in SF and before that, he was obviously someone we already knew about, right? Like, open AI, you know.

So, yeah, I mean, of course, now... The whole world is using Chan GPT, so there's an increased level of fame. But back then, he was a celebrity for me as well.

This time also, you were living with your parents in a shared apartment because he was a PhD student, right? Because my parents were PhD students. Yeah, my dad was a PhD student.

So how was that like for her? So that was in Miami, right? After HFZero, I ended up raising some money. So I would go to these different conferences, you know, San Francisco, Boston. I would crash at friends places and stuff.

And that's how attendees conferences meet people. And your parents were always encouraging, right? Or you had to fight to get to convince them? No, definitely. I think once I kind of gave them the proof points, you know, that I've gotten somewhere, I have so and so funding, or somebody is vouching for the growth of whatever it is I'm doing.

They were pretty, pretty encouraging. And in Business Insider, it was written that at the Tech Week Miami, like, You accidentally got introduced and funded. Was that true? Accidentally?

I mean, not exactly accidentally. It was... I mean, I didn't plan to start a company, right?

I was just there because Hackathon and meeting Michael Samen. So, at HF0, I went in with about 10, 11-ish ideas generally that I wanted. HF0 is BC, right? HF0 is the three-month residency that I did after the Hackathons.

residency yes so it was um so back in capital was running hf0 it was three months founders living in one house and building a product yeah and they all get funded okay so in that in that incubator kind of yeah incubator so once you got the funding how much was the first funding that was actually hf0 now they're investing a million each per founder back then they were investing about 25k wow yeah wow we've grown a long way yeah you Okay, so from here you started living in an incubator. Yeah. Started living in an incubator.

I ended up moving to virtual high school. So my entire education was now asynchronous and online. How did you manage? Like, you know, 11th, 12th, you talk to a JEE student, they cannot even finish assignments and you were building stuff.

I mean, yeah, at that point, my parents have had this conversation with me multiple times, how different it would have been, 11th and 12th, personally, for me, if I was in India versus here. And yeah, I would probably be a JEE grinder if I wasn't here. But yeah, at that point, I think I consider... You know, the opportunities, I'm very grateful for the opportunities I got at that time.

I think I 11th and 12th. Very interesting time. I think even compared to the average high school over here, right, they're all just like kind of on this track, getting into a good school.

And I was just in a completely different space, you know, working, living with people who are 10 years older than me. So yeah, I think I was the second youngest person was 18 and then 20, 22 and so on. And you were 16, right? I was 14 in the incubator.

In the incubator. Okay. So 14 to 16, you were in high school. Yeah. In incubator, managing high school.

And at the time, you were not thinking about college, right? You were thinking about building and finishing assignments. Yeah. That's how you managed. Okay.

So you've graduated high school at like, you're now 17. Yeah, now I'm 17. So you graduated high school. 18. Sorry, I just turned 18. Okay. So you graduated high school last year at the age of 16, 17. Yeah. So- Now, why college?

Like, you said you were inspired by Michael Sammons, right? Who never went to college. Yeah, who never went to college.

So, why college for you? I think somewhere along the line, I realized that being a solo founder is not great. You definitely need somebody to share that burden with. You need somebody to help you along the journey.

And being in college, college is one of the best places to find a co-founder. You know, I used to read Paul Graham's essays. And I think it's something... said something like 40% of founders meet their co-founders in college. Yeah.

And that is the biggest, you know, kind of, the place with the highest number of, how do I say this? Founder mode. Yeah, exactly.

We find people here, especially at a college like Georgia Tech, I think the startup ecosystem is growing. And I'm finding, I'm meeting people who have actually done startups in high school as well over here. So a lot of builders, a lot of talented, you know, engineers.

So I think it's one of the best places too. So before now, you were building everything alone? Yeah.

Team of one? Really? I had engineers, of course, you know, on the team, employees and stuff, but I was the solo founder. So how big was your team when you were like in high school?

About four people. So all were in India? Remote, remote. Some people were in Canada.

One of them was in LA. One in India. Oh, and how did they feel about working for a teenager? I think they don't really care if they get paid. At the end of the day, that's really how I describe it.

A lot of people ask me this. How did you work with people who are so much older than you? No one really cares if they get paid.

So in your entire journey, it never happened that... I don't talk to kids That's a comment That's never happened That's never happened I think if you talk to like maybe old people right who are like kind of 60 plusish So you'll feel a little bit of a divide you know They feel like what is this kid going to do That type of thing they have that yeah but generally I've never felt that You said that you have lived in incubators but you still plan to go to college Isn't that a very expensive investment like 40k a year like especially because you're out of state right you came from florida and like you if you could find co-founders at an incubator wouldn't that be cheaper so uh i would have not gone to college if not for the scholarship i was i ended up getting a scholarship from master's own uh so that was the master's own foundation uh master yoshi son founder of softbank he has his own foundation Colmar Stone and it basically funds young people under the age of 28 and their education in the parts of their building. It's only for the US probably, right?

No, it's international. Really? Yeah. So even if you're an international student coming from India, you could have been a legend.

Exactly. That is what I... Like, I was an Indian citizen.

There was no way for me to get these sorts of scholarships for US permanent residents or citizens or anything. So it was Coca-Cola's, you know, that... I think there's a scholarship for Coca-Cola. There's the Bill Gates Foundation. I was looking for all these options, nothing was there for me.

So I ended up founding Masterzone. That is incredible. That's, I think, all international students need to hear. I definitely, I agree.

Masterzone Foundation. How did you find? I, so through a friend, actually.

So I used to go to, as I... told you I was going to San Francisco and Boston, these places. I ended up meeting somebody at Berkeley who is a student from Sri Lanka and he got this scholarship as well. Awesome. Now, in your journey, have you ever been like broke?

Like, how did you fund the first biggest... Have I been broke? Yeah, in this journey.

Like, how did you fix that? Like, you need capital to do at least something. For example, if you want to go to maybe San Francisco, you need $500 for the flight.

I mean, I ended up raising money a little bit after. So, you know, HFZO happened and then I ended up getting some money from ODX on deck. And then AngelList as well, 2023, a couple more angel investors.

So, that's what initially helped me out. So, once you got the first funding, you started fighting and then it happened. Awesome.

So, now let's get to know more about your project. So, like for even for let's say a five-year-old kid can... can you walk them through Delve AI? Yeah, of course.

So, I was actually just reading a book on this, Information and Society. And it was talking about how our entire world is, you know, information is growing at a rapid speed. And for a five-year-old kid, how would I explain this?

Literally, for a five-year-old kid. Five to ten, you know. Like, think as smart as you, when you were ten.

Think that way. I see. Because we want to make those ten-year-olds... old kids watching as well as smart as you are okay imagine this um you're really interested in a specific topic you want to know why the sky is blue um and you go to a library and you see 5 000 books there and you don't know which one to start reading to find the answer to your question how would you do that how would you approach that problem that's what dell solves so and you have a lot of competition in this space right you have Like similar projects could be consensus, right?

That's a GPT, in chat GPT, like you can, you have questions about five research papers. Right. You can ask questions.

Right. You have perplexity now. So how do you think you are unique? Right. I think search by itself right now looks very different than what it will look like in another 10 to 15 years.

I'm looking long term, specifically just focusing on like the backend to improve search. I think both perplexity and consensus are built for like the broader consumer. you know there aren't a lot of tools that are built for enterprises or very specific use cases so we're focusing on consultants right example with anna specific use case um yeah of course i think consultants specifically um like they have like five or ten different uh data bases they have data in right they have like sales force have spot blah blah blah um then there's also public data they have to do research on like specific topic it's very disparate is what i'm trying to say um And that's why I made the analogy with the library, because they are that one specific customer segment for which information is so desperately stored in a lot of different places.

For researchers, it is still consolidated in journals and research databases and stuff. But for consultants, it's a lot harder to pinpoint even what segment to look into to find the information that they're looking for. So that seems like a very complicated project. Like you have information coming from five different sources because Internet is organized thanks to.

you know google search searching it but your personal company enterprise data which you mentioned is not organized right so how did you build up how did you build elf how did you solve this big problem right i mean we're still solving this problem right we are young startup which is 2.5 years old um but yeah at the moment uh initially two years like about for one and a half years of the initial uh journey uh we were building for researchers um and i quickly figured out that that problem is you know I think researchers as a market is a very tough market to crack and I'll admit this for anybody who's thinking about building for researchers they do not have the incentive to make their processes efficient internally because why think about it okay researchers is that one like researchers are that one specific industry where quality assurance is not a paid is not a paid position peer reviewers don't get paid to do the work that they do there's a lot of things that go wrong you can't treat researchers as like a regular industry that would invest in money to make things you know more productive why can you decide like yeah because maybe someone who hasn't done much research yes why they don't have peer review why they don't have qa yeah um they have peer reviewers but the peer reviewers are just just aren't paid it's not like that's just not how research research works um in research you don't pitch to get to get money you just write grant proposals and you have to stick to the grant proposal and you can't sway from it even if you think that there's something more novel out there. Like the way that research works is very different is what I'm trying to say. And for someone like Enterprise, I think it's a lot easier. to scale and build for them because they have like a very clear objective which is to just make the process more efficient reduce gray work and have as much output for the least input so research is a complex problem to solve because a lot of things are in development they're not proven yet and they are not because they can be proven because they are not used by hundreds of people so java ship jaldikar if things are shipped they are proven yeah but those things are not proven yet so that's why they are in gray area Yeah, I think that's one, like, I mean, obviously, research, it's R&D, it is still, there's a lot of things that are unknown, a lot of unknown variables there. So that's one thing.

And of course, the way that the budgeting is done, you know, grant proposals, even the top research labs, they don't have a lot of just extra money around there, you know, to invest into other stuff. So I personally, like, this is what I thought that maybe I should start by focusing on just the top universities, and then I can create. products for like the broader consumer, broader market base. And I would just go to offices of Harvard Medical School professors or MIT professors and pitch it to them straight up. But I think people, I think finances were a big issue, even at those labs which have a lot of resources to invest into things that are outside of directly the scope of whatever it is that they're developing or investigating.

So you found a big problem and now you're finding niche categories like university. Yeah. yeah okay so so how many customers you have so far um so we did about uh six enterprise pilots and currently we're about 250 plus individual users um and 250 plus 250 plus yeah and there's 50k 15k on the wait list that is still yet to have access to the product i have applied as well okay soon i will i will keep in the priority hi uh this is a demo to our current APIs, clustering, and key sections. And so I've run a query on COVID-19 sort of on the entire corpus of PubMed documents. And the source sort of serves as an indexer for all medical-related documents.

And three or four customers are actually using it at the moment. And so I've asked a question around autoimmune response to COVID-19 patients when they specifically have Down syndrome. And so to index the documents that specifically talk about that relationship, while also understanding all the underlying topics and then doing second degree analysis, that's sort of what I'm going to show you. And so it's come up with around 6,000 documents, just under 6,000 documents for this particular query, close to 200 themes and topics and sort of clustered all these documents out in these little rectangles, bigger sections, bigger rectangles and going smaller depending on. how many times that particular theme occurs in that entire corpus.

And so I have all the documents in the scope on the right. Let's say I want to understand maybe thrombosis or biomarkers or whatever better. Under that particular query, I can click Analyze Second Time, and then it runs the same theme and topic analysis on that particular query under the 6,000 documents. And so you have a new tree map of all those topics.

under that one sub-topic will look like. And so yeah, you have a second degree analysis and that's what the clustering looks like. The second part is key sections, which is once you have a document that you wanna go deep under. So what I previously showed you is like a lateral view of the entire domain.

Now, once you have one document that you think is like very apt or you want to ask for, a specific question to, you just upload it here or even just give sort of like an open source link and then you can do Q&A on it. You can do one question on multiple files so you sort of save time when let's say for example somebody's doing a literature review, so they get all the paragraphs that are semantically associated with that query that directly answer that question in all of the documents that you uploaded. And so this is my account on the platforms I can show you previous queries that I've run. So it sort of saves everything in stored results. And yeah, I can sort of just go here and see what that looks like when you run a particular query on a doc.

And so, yeah, so it's extracted those paragraphs that answer that question directly. My question was how good are neural networks at writing legal text? And that's the PDF that I wanted to ask from and sort of created.

a coagulation of all the paragraphs that directly answer the question. And so, yeah, that's what the product looks like. You push your product to product hunt as well.

So I think you were actively looking at product hunt. Did you build this to go after the hype or you were genuinely interested? I think...

Or both. So you're mentioning AI specifically? Yeah, specifically because everyone is building stuff since 2022. So I started about the end of 2021 and everybody in my cohort out of... 11 companies right, the other 10 companies they were all crypto.

I was the only AI company back then. So I like to think that I was there before the hype cycle came on. Yeah, so I think that's definitely there. And it was a core problem because I used to do research in high school and then that's where the problem kind of stemmed from.

That's a great way to look at it. Like, you know, I think the Perplexity CEO just posted as well. Like, what is the top product of that year and what was the hype?

VCs hyped crypto, but the top product was something else. You know, that shows that if you just go after the hype, maybe you might not be able to make it. But your hype may come later.

Right, right, right. I definitely agree with that. And mine came very soon, right after I launched. But yeah, at the moment, I wasn't really seeing a lot of AI companies. Product Hunt was like a lot of crypto companies, you know, increase your savings by, you know, a bigger APR. by investing it in crypto and stuff like that.

So a lot of ideas in that field, decentralized finance. And then AI got on pretty soon after that. And what were the biggest lows and highs of your journey? I think, let's see.

I think biggest highs, there were highs and lows and then highs again and lows again. Highs definitely, you know, initially getting that, being in HL0, pushing to product hunt. And I will say this about HF0, it was probably a very productive time. As for an incubator, for any other hacker house that I've lived in, I've never, you know, kind of felt this sort of like infectious energy going on, you know, just like putting hours and hours into the project, bouncing ideas off of each other and just working hard. I would see people working until 6am every morning in HF0.

So that would, as a 14 year old, you know, seeing that sort of environment, it pushed me as well to do the best I could. There was no excuse because I was the youngest or whatever. I had to do good. So, yeah, I think that was one of the moments I did a lot of work, product hand happened, ODX happened, getting investment from them.

And then, of course, the tough times came on. I had, you know, an engineer. We were working on a product. We tried to raise funding towards the end of 2022, but then there was a little bit of a recession.

So it was harder to raise money. So we weren't able to close that round. That was. kind of a low point I would imagine.

And then, 2023-ish, my first hire ended up not working out, we had to let go. But we ended up continuing the cycle of growth, that was something that I kept consistently working on. And then, of course, the media impressions happened, a lot of people came on board, we got a lot of attention towards our product and now we're consistently growing. and I have a team of overseas engineers that are helping me work on the product.

So it is sustainably growing. Biggest low, tell us. Maybe you thought that you took the wrong decision.

Maybe the startup is going nowhere. Let's see. Yeah, I think we were going to close an enterprise contract in early 2023. University or company?

It was a university, I remember. And at the same time, I was about to... I was pitching to VCs for a seed round as well. That contract ended up falling through at the very last moment and so did the VC funding. And that was one of the lowest points, I imagine, of my entire time building this company.

And three, four months of just, like, you know, kind of not even having the motivation to, like, get started with the day. A lot of, it was quite difficult, I think back to that time. But, I mean, of course, my parents, they were my biggest support system.

And I made a lot of friends through the teenage startup community as well, through different events. So, you know, that helped me push through. So, let's say if you have to look back, you're giving advice to a 14-year-old Pranjali, what would you say?

I think I would just say, do what excites and don't be afraid. of, you know, what may happen. Just keep doing whatever it is that, you know, excites you the most, that you think. Yeah. I think that's literally in three words, do what excites.

Amazing. Let's come back to the Dell VI, how internally, how the tech works. So can you break down the engineering? It has RAG systems.

At what time do you use AI? So we have, we're using proprietary models, you know, fine tuning Lama and Mistral and using that to do RAG on very specific use cases. So that's the AI part.

And then, of course, as the user uses it more, they're able to read the responses and we use that feedback to improve the model over time. Lama and Mistral. So why Lama and Mistral? How did you choose those?

Yeah, I think because Lama and Mistral, of course, you know, launched by Facebook, open source model. I tweaked around with it, ran a few fine-tuning experiments just on my own, on like biomedical Q&A, biomedical question and answering tests. And it was doing really well.

And I was just surprised that Lama has that sort of ability to, you know, I gave it about 100 or 200 odd examples and it was able to generalize and answer questions. So, you know, I think that's another thing that I would have good advice to young founders who are building in AI. Then build your own conviction on things, read papers and actually do the background research whenever you're deciding to, you know, take a step in your product or whatever. Do your own tests. For example, there is a very popular tool called HeyGen.

I was making videos, just trying if HeyGen is for me or not. You give a voiceover, it does the lip sync. And in my same clothes, which I've trained them all before, it does it perfectly.

But then I realized it's not fun anymore. So I think people have to go and try whatever the cool tools are there. Just don't listen to someone's feedback. For sure.

Build your own conviction. Absolutely. So you got the conviction and what else is going in the tool?

Yeah. So, RAG, of course, we have RAG there as well. I've been experimenting with hierarchical RAG. That's what it's called.

So, I used to use Pinecone initially. I ended up looking at the documentation of Pinecone. Sorry.

Pinecone is a vector database. So, it's used for semantic search. You know, store data, textual data. Like vector database.

Yeah, basically. And they have a system called hierarchical embeddings, where basically you use... The concept of knowledge graphs to make search easier and more efficient. So imagine this, instead of just having your vector data embeddings just disparately stored, you cluster them first, like you create sub-communities and you create communities on top of them and like a kind of hierarchical system essentially. So we use that to sort of make a lot of data.

Data essentially is dense. when you're thinking about doing research or consulting or whatever it is so um higher he collected database uh for rag um proprietary models and then of course encryption and templating to protect um you know customer data impressive and overall how long if a student watching this wants to build this how long it takes and if you have to rebuild it how would it how much it would take oh If I was to rebuild it, I would definitely not take as much time because now I have all the experience, you know, like, at least the knowledge that I didn't have before. A lot of trial and error happened, you know, before I figured out that these were the tools to use to build something. And the thing is that I think when you're solving a complicated problem, they always say just build an MVP, but build a high quality MVP. Don't build just like a regular, you know, like...

hardcore yeah essentially um like i i used to use a cosine similarity i think um initially and ended up realizing that that would go out very fast um cosine similarity is essentially another technique in virtual databases to find similar like similar text to do search and um yeah it's an it's an easy way to do search but um you always have to think about what's the best what's the best uh thing to do for your user how can you deliver value to them in the most like less time and for that you may have to do more research and not just take you know the first solution that shows up which in my case was initially cosine similarity and then i did research and found out about hierarchical embeddings and all of that i see so trying different solutions is very important so figured out that cosine similarity is not the best way hierarchical one hierarchical embeddings i think it's just doing a lot of research initially i didn't really look into the back end of machine learning and do all that study work because you know i was just focused on building talking to customers and just doing the tasks and i think taking out time in like recently i've been taking out more time to actually look into the back end of uh of the product do research on by myself on machine learning and um yeah I think that's helping me definitely. I see. An interesting question could be, what are the technical expertise you have versus don't have?

Like, what are the coding knowledge you have in AI, machine learning, or maybe DSA, and what you don't have? I think it's easy. It's important to accept what you know versus not know. Yeah, yeah, 100%. So can you break it down?

What do you know and what you don't know? Right now? Yeah. And what are you learning, you know? Yeah.

I think over the years of building Delve, I've learned a lot about full stack. And now I'm definitely learning more about machine learning. And as for what I don't know, you never completely know what you don't know.

That's always there. But yeah, I think for now, I'm going to continue on this exploratory journey of NLP and ML and then figure out what next to learn from there. So what do you think is the best way for kids to learn programming, machine learning, AI? Yeah. I mean for me it was getting a little bit of head start initially with you know comparative programming and DSA and all that and then eventually getting into building my own projects.

For young kids I would honestly I would encourage them to look into building their own projects and then first of all starting with like a very basic language, building something simple and then think about what would they would want to add on top of it and then learning new languages to do that. So it's like utility based programming like um what's what's the word? Inclement?

Bottom up. Bottom up? Yeah. Example? Example?

Uh let's see. Mujhe game banane hai. Uh I create like the most like basic version of it then I decide I want to add something on top of it and then eventually I just keep learning things to make whatever whatever it is that I want.

I think that's the that's one of the best ways to learn stuff I think. I see do what you enjoy and build on top of it like slowly add more features make it more complex because it serves as motivation to like you know keep going um i think a lot of people learn for the sake of getting a job but you know eventually that's what ends up happening but if you learn uh if you keep what you enjoy it you know at the front of the forefront then you know you don't need external motivation to push you now coming back to college degree which you first thought you will not pursue inspired by yeah what are the things you're liking about your degree georgia tech experience so far and do you still think it's worth it um are you able to find your co-founder yeah it's only been the first semester to start to be exchanged i definitely met a lot of interesting people you know you um chris klaus and i've gotten like access to the startup community over here so yeah i think that's been a big support currently i'm still building on my own i am going to look out for my co-founder i'm just building actually I have met a lot of cool people here and been going to hackathons with them. So I'm sure, you know, soon, soon I will find someone who's the right fit. I remember Startup Exchange Competition at Georgia Tech. There were three winners.

All three were Indian. And they're also Indian citizens, not even Indian Americans. Yeah.

All three were Indian citizens. Yeah. So can you share, like, what is something in us now that we are on the forefront of everything?

Like, we are, for example... whether it's like top CEOs of Silicon Valley. Yeah. We are there. Top, top, you know, entrepreneurship competition in the US.

We are there. Maybe it's a small sample size I have, but what is in us? That's a good question. I think we have, I think we have a lot of fire inside us.

We have a lot of ambition generally. I think. I mean there's a little bit of a sad duality to this as well. I think because we come from, a lot of people come from middle class backgrounds in India, I did as well.

So sometimes it is about proving something to someone and it's not necessarily internally there. But I like to believe that us as a country, we have a lot of fire in us and we have the ability to work hard. That's why we're also, you know, the... I'm pretty sure the highest earning demographic in the US as well.

Like you know post 120k, next is straight 90. Yeah exactly, look at that. Yeah so on top of that I would like to share that I think we are really good at finding out the hype, things that work like for example a lot of students are maybe not going to get jobs as easily with a college degree they used to. So now the next good thing you can do is to catch the hype of building stuff and they are good and in finding opportunities wherever they are possible yeah and that's how you see so many founders yeah i agree about being resourceful I think that is the key. That's the key.

It's not what you have, it's what you do with what you have. And I think that's what my parents also instilled in me, I think, as a kid. Being resourceful is one of the biggest tools that you have over here. From finding hackathons to go to, to finding scholarships, to basically figuring out where it is that I can get what I want. Who are the most inspiring founders?

We think a lot about the CEO of AMD, Lisa Su. She's definitely one of those very, very inspirational people now that I think about it. Initially, you know, as a young kid, I never really thought about, you know, sort of the gender aspect of building in tech.

Now that I'm getting older, I'm in college, I see that more clearly. And this is why I respect her a lot. Of course, Sam Altman, for what he's built, he's leading, you know, the change in Gen AI. Who else?

who is the ex-CEO of GitHub, Nat Friedman. Nat Friedman is definitely somebody that I look up to. Yeah, these are three names that come to my mind.

Why? Can you share the reason too? Yeah, of course.

I used to look up to his knowledge. Oh, it's Patrick Collison, CEO of Stripe. And Nat Friedman and Patrick Collison, both of them specifically because I spend a lot of time on their websites. I read a lot about, you know, advice, things to do before you're 20 on Patrick Olson's website, on Nat Friedman's website. And then he has, if you notice something, a lot of people in tech have very simple websites, you know, where they have their insights and all that.

Yeah, like Patrick. And yeah, I think just having that distilled, just the knowledge that they've gained throughout the years, having access to that is something that I really, I really, you know, read them, read it a lot. Yeah. And, you know, you got...

The top advisors on your board, for example, Lucy Gao, who founded of Scale.ai, who got kicked out. That's a sad story. So I think, what do you think about networking? At a young age, networking with top folks, taking them as advisors, networking with Sam Altman, before the AI hype. What do you think about networking?

What do I think about networking? How do you partner with these people? Every single place where I've been able to meet these people, anybody could have gone. That is something interesting to pick. Future Forum, the conference that I ended up meeting Sam Alpinen or Miami Hackey, we were going to meet Lucy.

Anybody could have gotten in. But it's about taking the initiative to find those things and go there. And I think sometimes it's not just about meeting somebody big. It's about finding whatever it is that you enjoy and finding those niche players that you really want to connect to.

For me, they were these people. But yeah, I never planned to meet Sam Altman, yeah, planned to meet Lucy Goh. I just ended up being there because I was interested in the topics and then that's how it happened. And then you showed interest to them and they cared about you. Yeah, I think Lucy Goh, that definitely happened because of, you know, HFZero.

I was building stuff, I had some good ideas and, you know, that didn't, however that happened, she saw value in what we were building, so yeah, that's how that happened. And for Sam Altman, it was just... asking him a random question and wasn't necessarily pitching my company to him um but even that that interaction to date you know it's quite special to me the big question a lot of founders are on o1 visa but you are a student are you on f1 visa which visa are you and how you got into it yeah um so i'm going to an o1 visa um initially i was dependent of my parents my um my parents were here for the for the phd so i was not dependent of an you know for student which was f2 then moved to h4 which is dependent of an h1b And then through Delve, I applied for O1.

And that's how. How long the process took? So I took a break in the middle of the process while the legal stuff was going on because I was more invested in like, you know, like deals and was just really busy in general. So I put a pause on it. But if you just do it in one go, it should not take more than like two, two and a half, three months.

Really? That fast? It is pretty fast. It has a two week turnaround from sending the application to when you hear back from them.

That's amazing. You apply for Express? I apply for Lexstra, yeah. Awesome. And like on O1 now, you have freedom to do internships if you want to.

Yeah, yeah. So whatever you want to do in your college journey as a founder, you have freedom to. It's basically an employment visa. Yeah. Like any advice you have for O1?

Yeah, so there's some companies that are new startups actually that are serving founders that want to do O1. One of them is, I don't remember the full name, but it started Plymouth. And...

I use LegalPad. It's a startup that ended up getting acquired by Deal, another company. But yeah, LegalPad helped me. And the average cost comes out around 2K, 10K-ish. 10K-ish, yes.

Yeah, yeah, yeah. For O1. That's pretty good because like your entire immigration journey is straightforward after that. And even as an Indian, after O1, you get green card like within few years. Yeah, yeah, yeah.

So that's amazing. Yeah, yeah. Actually, there's one, someone I know, Sigil Nguyen.

He's another young... he's pretty famous in the young founder community he has this community and then he gives out hoodies to people who are recent recipients of the O1 it's called the Extraordinary Club I think something something like that but yeah there's a lot of resources on his website I can link share you sure thank you so much for sharing everything in detail and you know inspiring all the kids now to build at a young age and I hope people take away how to go to college fully funded and how to build stuff thank you so much again of course

Transcript for:Pranjali: From Student to Startup Innovator

Transcript for:
Pranjali: From Student to Startup Innovator