so Neema I [Music] think hey everybody it's uh Jay Myers here CEO of uh of enen and we're here with another project uh podcast this is better than Joe Rogan so uh uh stick in with us here I'm joined today by Neema zamani who is co-founder and CTO of a called cobionic uh so Nema tell us a little bit about yourself uh and cobionic and how this uh how this organization got off the ground yeah sounds uh fantastic Jason again thank you so much for kind of inviting us to the podcast uh so the story kind of started you know me and my other co-founder his name is team blell we've been kind of friends uh uh ever since undergrad you know so we know each other for almost 15 16 years now and uh uh so we've been always wanting to create a company right I mean we've had probably tens of ideas to kind of start you know and every now and then you know we had a new kind of uh uh uh idea that we wanted to kind of uh uh start a company on uh but uh unfortunately time never worked out you know or there were different situations that kind of prevented us from uh starting a company so what uh uh uh happened during Co both my other co- fer and I had this opportunity to kind of uh uh uh see how we can help uh the situation with Co so we were like okay you know what what's the what's the main problem people are having right now and that was vaccination uh so at the time we were like okay you know what how can we make this more streamlined how can we make it easy I mean it obviously was a huge problem that doctors and nurses had to kind of leave what they were doing uh uh uh uh and then uh go and vaccinate people so we we had surgeons that had to stop surgery and vaccinating people we had it was it was a complete uh uh mess you know for the sake of a you know better word but uh so what what we kind of did was uh uh uh said you know what let's bring Ai and Robotics together you know so let's let's use robots to kind of vaccinate people so that was our kind of a first kind of a proof of concept initially we got a lot of push back on it you know because people don't like the idea of robots and needles and you know I mean it's pretty it's pretty scary imagine a robot coming at you with a needle but we decided to go with a needleless solution uh uh so there were no needles you know so it was all pneumatically injected uh and uh we had some clever designs for that you know and that idea kind of took kind of a life of its own for a couple of months in the Press you know we were featured in BBC uh you know and Bloomberg you know Fast Company other places say you know uh uh and uh uh so that was the start of the company that's how we got our initial fund you know uh to kind of try to bring the state-ofthe-art Robotics and artificial intelligence to the medical space uh uh so obviously Co died down you know and the interest in vaccination went down with it you know uh uh uh so uh we kind of shifted our Focus from vaccination uh to Ultram so so right now the robot you see behind us is the robot that is funny enough partially funded by engine so we kind of made this robot from uh uh scratch right so the robot is a medical grade robot uh it's got pretty much all the states of uh Dr technology in it uh we are using the you know the best control systems the best servos uh the motion technology you know the gearing uh the sensors you know there are about 10 cameras in the robot so there's a lot of Technology packed into that robot uh if I'm not mistaken we are uh kind of the only the third company in the world that's make that makes uh medical collaborative robots you know the other two being Kuka and kinobo uh who who are also based in Montreal so they're also a Canadian company and uh uh so uh it's it's a very small Market it's an untapped market and I kind of feel like like it's a necessary Market you know because uh as the population ages as there's more need for medical staff uh uh uh the only kind of way to solve it is to augment and enhance the uh uh medical staff right now with more technology and and part of that is going to be kind of Robotics so I'll stop right there you know and uh yeah no that's uh that's great um and a lot to kind of unpack there uh too so going from kind of an idea about how to work with uh with one of your closest friends now to you know sort of one of the three leading uh medical robot companies in the world is not uh not you know a slight uh it's something to to Really um uh really congratul congratulate you guys about but uh uh it's a real uh a real achievement here and as you say this is not uh I mean it's not just uh any robot company it's a a cobot company yes which I think is really important too but what's your what's the basic Vision behind uh behind what you're doing because it you know it can be used to as you say to deliver vaccinations or uh or ultrasound uh here but clearly a lot of other applications too what's the what's the kind of your overriding vision of uh of the technology so so I mean there's no question you know especially with the advancement in AI there's going to be uh there's going to be this new explosion of robotic applications right I mean what uh has classically held robotics back was the AI portion of it you know right uh I mean obviously robots have existed for well over 30 years you know uh in kind of different forms and but but what has made them uh uh not uh very uh kind of useful in everyday tasks uh was was the uh uh uh lack of AI uh uh because in everyday environments you know if you're like whether the robots like Philippine burgers or whether it's doing ultrasound anytime the robot is in everyday environment or dealing with humans uh it's a chaotic environment right I mean things are not ordered you're not in a factory you know so usually factories everything is ordered you know uh you know exactly where the part's going to be you know exactly how to pick it up you know exactly what the part is uh but in everyday environments things are very chaotic and in order for robots to do anything useful they have to kind of think under feet you know so they have to kind of be able to make judgments and decisions on the flag and the way to do that is using Ai and also kind of state of the-art computing uh so so there are multiple Technologies I kind of feel like is converging down uh to kind of a single point and that's why there's going to be this robotic and AI revolution in the next uh 10 to 15 years you know uh similar to how the inter internet uh got us started and one of them is uh uh the exponential increase in computation power very cheaply right so all the AI for example we are running on the robot behind me is being done on the edge or it's been done right uh uh right on the robot uh uh uh so it doesn't have to be connected to the cloud it can just think on its own if anything happens can react to it and that is crucial you know for any robot cob application that's uh that's going to uh deal with the humans that's going to deal in a chaotic environment the other thing is that the uh uh the other technology that's converging is the decrease in the cost of uh actuation and sensors so some of these sensors existed before but they would cost significantly more than it cost today uh so so like for example One camera could be like 10 times the cost of a camera today you know with the same level of you know technology and capability so so there's a huge decrease in the cost of the uh the sensors and actuators and the third would be the advancement in Ai and this is this is what's happening in the past 10 years or so uh uh that that's kind of making this all possible so it's it's a convergence of multiple Technologies I and I think that's really important I mean we've got uh AI Computing sensors uh you do a lot of in materials uh material space too and then materials research uh around and of course on the medical medical side it's uh I've always said look you know people talk about machine learning and they tend to think just about AI I mean there's today everybody's focused on AI but you need to you need to integrate that even machine learning it's AI but you need a machine and so you know and it's how that uh how the total integrated solution is put together and uh and that's what you you guys were trying to do in the project so tell us a little bit about what you did with the uh the project that uh that we helped to support uh so so well the project we did with engine uh interestingly enough had nothing to do with the medical space uh so so we are also working on kind of the non-medical applications too so uh where can we apply the same technology uh that we are using for the medical side uh for non-medical applications corre and as as you mentioned rightly uh uh Jason one of the applications really interested in is in Material Science and using robots to kind of enhance uh or increase the innovation in Material Science right so classically Material Science has been uh kind of very uh uh expensive and and the reason for that is the labor involved you know so it's it's a lot of testing lot of trial and error uh lot of people putting samples in a test tube and samples you know in a test machine and doing that for you know months or years uh uh to be able to kind of fine-tune the material uh so one of the one of the things we did is that we kind of had this project with a company called Universal matter and they kind of manufacture uh kind of uh uh uh they have a process that is highly efficient at manufacturing graphine okay so graphine has classically been very uh uh expensive to produce uh and there have been a couple of companies but this company can do it very cheaply and uh uh uh so what we try to do with this company is that to bring the robots to the company to not only help with the manufacturing of graphine but also uh uh create a kind of a Smart Lab if you will uh uh so we can take samples from the uh let's say the production line test them as they're being produced or help with the R&D side uh so that is uh where we see cobots uh kind of playing a huge role in the Material Science and Manufacturing uh to be able uh to go around the factory take samples put them in the test machines get the results back and so we only uh we didn't only uh uh uh Supply the robots we also help them with the database as well so there's a whole database in the background uh uh that uh kind of coordinates the samples and the results of the samples you know to uh uh uh uh kind of uh track uh the quality of the production or to help with the R&D site uh uh and and once we can create this kind of a streamlined system then the amount of uh test data is going to increase substantial uh uh the costs of obtaining test results are going to decrease and then once you have the volume of the test that you need then you can use Ai and machine learning again uh that looks at this database and try to okay say you know what these are the input these are the outputs you know so we we did XYZ to the material these are the results on how the material you know uh perform uh uh so we can have an AI trained on it uh to be able to optimize uh uh uh uh uh for the for different products you know you know it's a really good example of there's still all sorts of U uh economists mainly but people doing analysis about the impact that Ai and and automation is having on the workforce and uh most of those people uh conclude that the introduction of AI robotics automation processes are are replacing jobs uh but I I think it's you know what you've what you've been describing is a really good example of where it's just the opposite it's the fact that you're trying to you're trying to save labor because the job or the people may not be available for the jobs but it's it's also you know you're adding uh so much more value uh in this process and especially around the data analytics uh uh here as well it's you're adding a component or tasks that it's not just repetitive tasks it's tasks that uh that people could not do on their own at at least at the cost or in the timeline that uh um you know that uh they would have available so it's I think it's an excellent example of how cobots especially can be used to improve improve value improved productivity but tell us a little bit about some of the challenges uh they of using cobot systems uh here are there um are there regulatory challenges are there health and safety challenges when you're when you're dealing with with robots with equipment that are actually working alongside people uh so yeah so the definitely there there's a lot of reg uh uh so whether it's the industrial cobot or medical cobot uh there are different standards we have to meet uh so as we were speaking actually our robots being kind of tested uh in in a place close to marom you know uh uh uh for its uh electrical and mechanical safety uh so any any robot that goes to the medical uh uh that has to get into a medical application has to pass What's called the IC 60601 uh and that contains all the mechanical and electrical safety so there's a huge range of tests you have to do uh uh on to see the robot is safe it stops uh at a safe distance you know if anything goes wrong uh you know it doesn't kill anyone uh uh uh what if what if there's a short circuit in the system you know is it going to explode you know so there's a lot of different uh kind of tests uh uh uh uh that you have to do obviously and uh uh the other thing is the speed that the robot goes around humans you know so classically cobots go at 250 millimet per second speed around humans obviously even the robot behind me can go uh kind of a multiple of that number you know it can probably be like five six times faster uh but but when you're dealing with humans or in a kind of a a human environment you have to slow it down a lot so even even if the robot hits you uh it doesn't really do anything so so that's so there's a lot of uh uh uh real regulation and there's a lot of uh I would say uh uh rule good good rules of thumb or suggestions you know that you have to kind of keep in mind when you're kind of creating a product like that and at the end of the day let's say for the medical side uh given all these requirements and Regulatory it's kind of the fda's job right to kind of give the kind of the final stamp of approval you know uh so to speak to say you know what these are all the regulations but based on everything you know your robot is safe then we approve it uh and uh then there it goes for the for the automation side or for uh uh industrial or non-medical side the bar is lower for obvious reasons you know I mean medical is always a higher bar uh uh so almost certainly if you can make a robot that can meet the medical standards it will pretty much meet any other standard yeah and uh uh uh so yeah so so there are definitely regulations and I'm assuming there will be more regulation especially on the AI side that's going to come in uh down the pipe uh uh because right now the AI is a little bit over like a wild west you know no no one really knows how to regulate it uh even on the medical side you know uh they don't really know how to uh uh uh uh kind of control it in a sense that a classic software uh there is kind of Version Control you know so you have uh you know you know the software is version 1.00 you know it's very uh uh uh nicely defined but with AI how do you how do you control that you know what if you train it again what if there's more data input putting being put into the training data is that a different AI or do you have to repeat all the test which probably would be in feasible because every time you update the AI you have to start from scratch it would you know be somewhat impossible so there's a lot of am ambiguity in the AI sense uh uh on how the regulation is done uh because uh there is AI is a kind of an interesting uh U interesting field there is the there's a structure of the AI you know which is let's say is uh more uh uh nicely defined then the the question becomes the training of the AI you know or the or the data that goes into the AI that one is the one that's not very nicely defined you know and uh uh what kind of data are you putting it in you know how is it being trained you know how so those type of things are are a little bit more ambiguous yeah so that's uh that's where I say you know the future is going to go on the regulation of the AI and and there's a lot of other ethical aspect to AI as well you know for example is the AI biased you know like towards you know different you know uh uh racial groups or different you know gender sex or all that stuff you know like and that depends on the data you're giving it uh to train you know so if if for example you're only giving it a certain set of data to being trained it's going to be biased towards that uh kind of you know aspect so there's a lot of both social ethical and also uh technical stuff which I'm pretty sure there's going to be a lot of Regulation right now we are in the kind of a wild west you know so we can do anything we want but I'm I kind of get a feeling that a couple of years from now there's going to be a lot of you know strict regulation on that so Nea how did I mean there are lots of issues out there uh around Ai and and you you've spoken to a number of the key issues here how do you see those affecting your your application though I mean clearly the AI is continually learning and that that issue about uh uh about you know you're not going to have one version but a a Continuum of uh and of hopefully improved versions uh of this going forward but wouldn't uh at least on the industry application side a lot of the are there still some concerns around the the bias of the data uh or on the medical application side too bias of the data that you would be using or or how do these issues how would these issues play into your your particular uh applications of AI yeah so so there there are a couple of things that you can do to make sure that you know over time uh uh the bias is minimized uh one way is to always continuously have a kind of a feedback loop right uh so so it's not like you send an AI Into the Wild and you kind of wash your hands from it and then you know okay you know what let's see what happens so you continuously have to get feedback so let's say if an AI has some sort of a bias uh uh then you realize that okay you know what let's say the AI had the let's say the robot had to do some ultra sound on a person but for some reason the AI had a bias and couldn't detect the face of the person or the body of the person then then you have to uh focus on that side and try to kind of uh do retraining of the AI and kind of eliminate the bias so it's going to be a kind of like a game of wack and mo you know so you gota as they pop up uh you gotta you gota you know try to uh uh uh uh diminish the bias but it's going to be very hard to create a first AI that doesn't have any problem you know just just like a h just like a human being you know there's it's gonna be very hard to you know get get a get a person who's 40 years old you know and who who kind of hasn't made any mistakes you know so so it's part of it's kind of kind of part of learning process I guess you know uh uh uh but but you need to have that continuous feedback right uh so that's why it is a Continuum and that's why you can just send AI into the wild and then wash your hands from it I mean and and we can see this in every day right now too you know for example with what's happening with Google now or what's happening you know with other companies with their uh large language models or the you know the diffusion models and uh uh uh uh and those are huge companies and they still have to deal with this problem you know because because hum because someone's going to find a way to make your AI break to say something that you shouldn't or to or to draw a painting that it shouldn't you know and then then all of a sudden you're in trouble uh uh uh so so those are the stuff that uh uh will get better over time and uh uh uh one of the other thing is uh the data site uh so there's going to be uh uh a lot of uh uh uh more refined data so AI is only as good as the data you give it so one of the things for example we find out whether it's the material science side or the ultrasound side uh first of all there's tremendous lack of data uh like we can't find good ultrasound data you know or good because all the data for example maybe it's not labeled it's uh scattered you know so it's not we don't have one National Bank of ultrasound data every Clinic has its own hard drive basically of ultrasound data you know and and it's not labeled and it's you know it's it's a just a it's a big mess let's put that uh there are people who are trying to solve that problem so one of the ways we can make AI better is by not really focusing on the AI side but focusing on the data side uh because again AI is only as good as what you're putting into you know so if if you have a large language models that uh is and that that that was one of the cases initially uh with these large language models that were trained on Twitter you know well there's a lot of crazy stuff that people say on Twitter you know and and if you get your AI trained on those crazy stuff then it say crazy stuff you know uh so so so that's that's where uh I think I think the focus of the future is going to be yeah no that's that's good well listen we talked a lot about uh Ai and the technology uh but one thing I wanted to ask you too because you're doing some really neat things with indigenous communities maybe uh speak a little bit and this isn't of course around materials Discovery uh but some of the applications of uh of your medical cobot what's uh what are you doing out in sasan oh yeah so currently we are in collaboration uh with sashan and both University of saskat one and saskat one Institute of Technology uh uh uh the uh we have our kind of a lead uh uh clinician you know to do the clinical study his name is Dr iar mandz and he's a Order of Canada recipient who is working Robotics and tele robotics he's a very smart person very interested in the field he's very keen on helping people and uh so we ended up having this kind of opportunity to work with them uh uh on the uh in saskat one especially in the kind of the uh Native communities in saskat one you know which which which are a little bit remote you know uh uh most of the times they don't have access to the best Healthcare uh so to speak uh uh and what we want to do is that first of all we don't want we want to as as as Ai and Robotics Advance we want to bring as many uh groups and communities with us through this advancement you know because we don't want them to be left behind so one of the things that happened unfortunately with technology is that the rate of growth for technology so fast that uh uh uh a lot of groups fall behind right because uh you know we've seen this multiple times in history you know ever since Industrial Revolution or you know so there there have been multiple times in history where things have exploded but it only benefited very few and it's left many people behind you know uh even though Industrial Revolution in France happened still 300 300 years later in Africa people didn't benefit from that uh you know Industrial Revolution so to speak you know uh so we just want to make sure it doesn't happen with robotics and AI part of it is because we kind of feel like the the explosion or the advancement in Ai and Robotics is going to be a lot more impactful than even the Industrial Revolution I think it's going to change societies you know fundamentally and and we want to make sure everybody benefits from it uh uh uh so so the goal with the native communities in S casan is to bring uh uh uh this technology to their uh Community to make sure they get access to the state-of-the-art healthcare and to also train uh uh uh uh people in those communities to deal with AI and Robotics you know I think that's also very important you know to just kind of decimate the knowledge uh you know uh uh uh into those communities uh and yeah so ultimately the goal of hopefully all of us on this planet is to make the world a bit better you know and leave it better than we got it you know uh and hopefully hopefully this project with Cashman is going to do that no it's uh we came across it I think through some discussions with our uh Saskatchewan uh indigenous manufacturing Contracting uh cluster that we uh that we work with uh uh but then also through some of the uh the school programs there and there all everybody was talking about this company cobionic uh that uh I working with um indigenous communities and uh it was great to for us to be able to say yeah we know them and uh and we've uh we've worked them uh with them as well so uh Nea just uh just a couple of other questions here for you what uh uh first of all this is this is a uh a company that's gone from from startup through some pretty challenging uh challenging times with Co and uh uh and and then also now you know sort of going into some really Leading Edge applications uh here I know it hasn't been an easy uh easy process for you as uh as co-founder and uh uh and your partners too but tell us a little bit about some of the challenges you're facing and uh uh and of course the you know the big uh the big question I think everybody has is uh uh are you able to find the people uh here to work with you uh yeah so that's a that's a interesting question first of all it was it was pretty challenging you know even even right now we kind of work everybody in the company Works about 10 hours sometimes 12 hours a day you know I mean it's very hard to get the uh you know to meet the deadlines you know to get the technology uh uh out the door so it it is very challenging uh uh we are kind of fortunate that uh uh uh uh we are in the kind of the uh uh University of watero uh Network you know so it's easier to find Talent you know the from the University uh of watero or even like you know from University of Toronto both of these universities are very good at Robotics and AI uh uh so we we are looking that side but on the on some of the other stuff uh uh uh that's not purely engineering uh is more challenging to find out you know and we are somewhat as a peak employment so it's very hard to get good experience people you know who are kind of willing to work for a startup uh and who also are located in this area right okay so there are multiple criterias that have to be met and that that's hard you know to find the people who are already here for experience who are willing to let go of their let's say if they're working for Apple or Google or someone you know to let go of their Google job and come work for a startup that's very hard so so there there seems to be kind of a kind of a limit on uh who we can get or uh how we can get them uh but so what we found that you know it's easier to just get uh maybe more Junior people and train them up uh to be you know kind of to that level but that's I mean you know it's kind of out of necessity but but it's seems to be uh uh uh the strategy moving forward you know to get Junior people and train them up to be like expert at AI or expert at robotics which uh yeah uh so you kind of have to not only build a business as a startup you also have to build the uh uh kind of train the people to you gotta do both yeah how how many employees do you have now so right now we're about 10 people you know full-time uh we have some researchers and part-time people so all together we're about 50 team uh uh so we have like three Re researchers in ufd PhD students working on some uh kind of AI models you know uh uh we have some part-time people or uh helping with the manufacturing we have co-ops but all together we are 15 people but 10 people fulltime and I know you have uh big growth plans so what's uh what do you see is next in uh in terms of of uh applications for your current technology but also you know next stage of uh uh of growth for cobionic yeah so we are currently in the kind of middle of uh doing our next fundraising uh so either it's going to be a bridge round or a series a round you know depending on how the economy goes you know again uh uh this past three years only hasn't been hard on the on uh the covid side or the uh or the the physical aspect is also been very hard on the financial side too yeah because the markets are very unstable you know one day you know there is co and everything goes down the other day you know it comes back up it's very volatile and it's making kind of decision- making very hard you know uh because uh you know someday they say there's going to be a recession then they say you know there's not going to be a recession then and everybody as soon as they hear the word recession nobody wants to give money then when they hear there's no recession everybody wants to give money so it's it's it's somewhat bipolar you know environment we're kind of living in but the uh uh uh so so uh uh the goal for us first of all is to raise the next round uh H which is going to help us expand the manufacturing and growth you know so we want to have uh manufacturing here in Ontario uh build the robots assemble the robots test the robots you know I think I think it's very important to have the manufacturing in Canada uh and uh uh uh there's no excuse you know to go to China or other places you know especially if you're a robotics company m the whole point of robots is a labor solution you know uh so if you're going to China or other countries for labor then why you building a robot to begin with you know so so the whole reason you build a robot is to bring the manufacturing back in back in Canada or or North America so to speak you know and uh uh uh the other kind of uh uh kind of our aspiration or goal is to put the Codi which we call the robot in many places possible we want to put Codi in as many places as possible whether it's medical or non-medical side you know obviously the non-medical application uh we can generate Revenue much faster because the medical side has a much longer uh path to revenue right because it got to do the clinical studies you got to do the FDA submission you got to do so there usually takes like three to five years you know uh uh but the non-medical side or medical adjacent side maybe if you're doing some sort of a massage robot or if you're doing muscle Rehabilitation or assisted feeding even on the on the medical adjacent side uh those will have a much shorter path to revenue yeah I'm just uh trying to think about a massage robot going 250 millimeters a second that be finished at about 10 minutes well well that's a max speed you know it doesn't have it can go very slow you can it can be very gentle yeah no but but listen you've got you've made great strides and uh with two well a couple of applications on the uh on the medical side and uh and clearly one on the uh the industrial side too so I think I think being able to demonstrate the application is is probably putting you in a really good position for uh uh for additional additional funding and and financing uh there regardless of what round you go to yeah yeah uh again thank you so much and engine has been a great partner through uh through that process uh we probably maybe might not have been here if it wasn't for that funding from engine you know uh uh because again part of the robot was subsidized by by engine and we are very grateful uh and hopefully fingers crossed you know we can continue this kind of collaboration with again hopefully many more projects down the line yeah yeah no that's great and uh and we're really pleased to have supported you and look forward to sticking with you uh uh here I I think it's you know one thing to for us to be able to uh to fund projects like yours but we're also really interested in making sure that uh uh that they're commercially successful too so look forward to working with you Nema yeah thank you so much Jason and a pleasure and hope everybody on the podcast enjoyed it yeah thanks a lot I I hope we've met the Met uh the standards that uh that Joe broken one day one day that's great thanks a lot have a great day [Music] bye