323. Building Next Generation AI Based on 20 Years of Research w/ Manish Patel | Jiva.ai

0:00Hi, my name is Mani. I'm CEO, co-founder of Jeeva and we are building deep tech AI in the UK that's going to blast away LLM technology in the next few years.

0:12Amardeep Parmar:So you told me that it came out of your PhD, right? Yes. And you've been working and tinker on this for years and years. Yeah. Tell us that journey. Yeah. It's a bit crazy, right? So I actually started my PhD at around the time the Human Genome Project was finishing. So back then, machine learning was relatively new. I mean, obviously, it's been around for a while, but, you know, it actually being applied to real world stuff was a relatively new thing. And I was actually a geneticist at the time, molecular geneticist at undergrad. My lab skills were atrocious. None of my labs worked. I was terrible.

0:51I don't know why I did that subject. But I was a much better programmer and much better at maths than I was at the lab work. So I switched over to that and I did my master's and my PhD in that particular subject. And it started off very much as a how do you solve complex problems? How do you solve complex systems? And we kind of discovered a way in which we can do machine learning at a pretty grand scale. But in a way that enables you to get lots of different models to talk to one another, merge and fuse. and that's quite similar to the way your brain works. So fast forward 20 years, we kind of realized that actually we could probably use this for real world machine learning, real world AI, and this would probably be the basis for real world reasoning.

1:41And so that's what we've been working on under the hood. And I think we were a little bit early, if I'm honest, to the pie. So we ended up doing a lot of no-code AI and that kind of thing in the interim. But in actuality, the real thing that we're going inside the company is how to build super intelligent machines, which is why it's called Jeeva. And I think a lot of you are Asian. Listeners probably understand that word, but we'll let the listeners figure that one out.

2:07Amardeep Parmar:And it's a critical, right? Because you said where AI is now this big boom, right? You did it before it was cool. Yeah. And obviously in those early days, I imagine you had to try and convince people that like, oh, this AI thing is going to be a big deal and people might not have believed you. People didn't get it. One of the big investors that we had at the start was like this high net worth network. And I remember I had to do this massive IM. And I was told to demote the platform vision type of stuff in the favor of a very specific use case, which is prostate cancer, which we were doing at the time.

2:45and we had to do that specific use case just because people didn't get that you could do this big platform where you're doing things. So it's weird that it actually dictated the way we had to take the company because investors didn't get it. And so we had to tell this story in this particular way. We ended up getting into med tech and healthcare and people still associate us a lot with healthcare. Actually, we're not a healthcare company. But it's weird how you are kind of made to go in a certain direction and just because you're either too early or people just don't get what you're doing. And I get it.

3:21I understand why that has to happen. But if I was to do that again, I'd probably be a bit more stubborn about not going that way.

3:30Amardeep Parmar:And your original co-founder as well was a doctor, right? Yeah, he's a doctor. So that made sense, right? So before we even started, he joined the clinical entrepreneurship program with Tony Young. and it's incredible entrepreneurs there and it's great to see how they've all evolved.

3:50So the idea was, look, healthcare is right here. There's a really big upside. Healthcare is a big industry, right? Really big upside. Even if we weren't industry agnostic and we were just concentrating on healthcare, that's still a massive market to operate in. And we did it. We did a prostate cancer diagnostic. We did a liver disease diagnostic. We've done various other very high fidelity healthcare applications. And we're showing that we do that on the platform. But the reality hits when you actually try to make that real, right? So I could create a diagnostic all day long. But to commercialize that, that is a whole different story.

4:28You've got to go through our own regulatory thing. You've got to have a whole dedicated team that can do those kind of things. That's quite difficult to put together for a small company and be a platform as well. So we made the strategic choice of, okay, look, we can enable those things, stay generic in terms of our platform. So we make a big deal of the fact that when a company comes to us to build something, the IP they generate on the platform is theirs. So if you want to build a diagnostic, you can build it on the platform and that IP is yours. You can do whatever you like with it afterwards.

4:58But hopefully it stays with the platform and you run it in the platform and that's how we make our revenue.

5:02Amardeep Parmar:How did you make your first revenue? Was it in the same way or how did you? That's how we did it in the first place. It was a weird one. I still remember the first client. It was a diagnostic. And we realized that actually, although we were saying we were a platform in order, license generic, actually, again, reality hits. You're like, okay, there's still all these things that don't work. You've got to figure all these things out. That's going to be like 12 months worth of development. it um so we realized pretty quickly actually we've got to get people on the platform get people using it we've got to get feedback and actually start building that in rather than it just being a brainchild of us as a team and we're all technical people so we're probably not the right people to be to be building it um or to be building ux had us uh yeah that first revenue was it was difficult together but we got it we got it made it work and and and got it got it running it's but it was a learning process for sure.

5:58Amardeep Parmar:And you also said as well how it's a Welsh company, right? And a lot of the team is in Wales. So we're, yeah, well, we're actually, the founders are all in London. Most of the team is in Cardiff or Newport. That's actually, it was quite fortuitous. We got investment from Development Bank of Wales, and therefore you had to headquarter in Wales, which is all fine. That was during COVID, so it was a weird time anyway, and no one was kind of traveling anyway. And what we found was Cardiff, Newport, Bristol, those areas have an amazing tech talent, amazing for about half the price you get in Reading and London.

6:38And it just worked because it was during COVID. We did actually technically, we had an office, but no one was going in because of COVID. Everyone was working from home. And so we just grew in this remote kind of way. And we ended up getting some really amazing talent on the team. And we pack a big punch, right? We got like a quantum physicist PhD, we got a machine learning PhD, we've got an imaging PhD, like crazy, crazy clever people for what I think was, you know, got a little bang for our buck, basically, right? Young in their careers, really raring to go, still really excited to work on stuff.

7:16It's great. I mean, I love my team. They're They're a great bunch to work with.

7:20Amardeep Parmar:How were you able to attract that kind of talent? My stunning charm. No. I think it's a couple of things. I think because we started off in healthcare, actually, there was that draw that you're doing something. I was setting the draw for my CTO. It was you're doing something for a greater benefit than just money. And whilst we're technically industry agnostic, I think that draw is still there given that we still do a lot in healthcare and life sciences anyway. So yes, of course, we have to generate revenue. We have to create value in the company. The stuff that we're working on is so fundamental.

8:02Like if I told you you can use Jiva to either create a diagnostic or create a virtual private assistant for yourself, that would arrange your meetings. Same platform, two very different use cases. The productivity gain is awesome for you as a founder, for you as a company, whoever you might be. If it's a diagnostic, even more so because there's a long-term thing there as well. So I think that's actually what attracted people to the company. is just the long-term benefits of what we're doing. I think that's really important.

8:37Amardeep Parmar:How has your role changed over time as well, right? Because obviously you started this when you started it and then now you've hired a team and you're doing some different things. Yeah, so I started off as the CEO. Right at the beginning, I started off as the CTO. We didn't have a CEO. We hired a CEO. It didn't work out. And my chair was like, look, Manning, it's got to be you. And I said, no, I don't want to pitch. I would much rather be sitting behind a screen, at least at the time anyway. Much harder to be sitting. This I would find really hard. I'd find it hard to talk to someone on camera.

9:14That's not my, I'm not innately built like that. And I had to really force myself. At the time, we were just jumping onto the KQ Live Accelerator. I don't know that one. And so it's a accelerator program in Kings Cross just for life science companies, healthcare companies. It's a really good accelerator, I've got to say.

9:34Amardeep Parmar:Francis Carrick. Yeah, yeah. Francis Crick Institute, yeah, yeah. Carrick, I was there, footballer. You're not Carrick. I was made then, because it was happening during that time, I was made then to do the demo day, which was then to pitch to 200 or 250 odd people at the Crick. and I like I remember that day I was nervous for two two and a half weeks I was all I could think about was how I'm gonna be when I do it because I feel so you know like you know pitching in front of people it's hot um so yeah I had to try to make that transition to CEO and I gotta say that's even if you've been a CEO before like this is my second startup not I wasn't C2 and the other one you just don't learn enough right and then when you're a ceo is a whole different ballgame because everything just comes up to you whether it's good or bad or whatever it is everything you're the only person that knows everything and it's really difficult to keep your head screwed on and mentally okay having to like do that transition right of like there's gonna be plenty of people as well who will say like okay i'm an introvert i don't want to be pitching get on that kind of stuff.

10:44Amardeep Parmar:Was it just reps or how do you think you got the confidence? Yeah, there's a few techniques, I think. So the number one thing I learned was I'm not good at memorizing words. I'm not good at memorizing, I mean, not words, I'm not good at memorizing whole texts. I should have known that from English, like English lit, we have to memorize a sonnet. Got about a third of the way through it and started making it up. It was terrible. But I'm quite good at memorizing bullet points. So now I think it was just about technique right so if i do it i'm doing a pitch i generally just have some bullet points in the back of my mind that i've memorized and then i'll just ad-lib through it and i'm quite i've found i'm quite good at ad-libbing and it's just like a self-discovery process right i still get really nervous i still get really nervous at pitches last week when you were there uh i you know when i get up there i forget about it but up to that point i'm quite nervous and but i can't show it i've You've got to remain confident and look confident on the way to smile.

11:46Amardeep Parmar:I don't think there's any fix for that. You just kind of just bring and bear it and just get on and do it. I think that's it. I think in my head, something just clicks and says, okay, look, just do it. I find what helped me in the end is I just got so busy that I didn't have time to stress about it. Just stress about it. Yeah, it's like I need to just get something done. Okay, I'm on stage now. Your mind doesn't have the bandwidth to deal with that stress. Yeah, he's like, yeah. It's like, well, our demo day is in like two weeks. And I'm hosting it, but I haven't thought about anything I'm going to say yet.

12:15Amardeep Parmar:And even I'm doing the talk soon. It's like, I kind of almost in the way, like, if you keep putting yourself in a pretty situation, it's like, well, I did it last time. It was like, yeah, it should be fine. It should be fine. Yeah. Yeah. I think for me, it's, I still, I need a framework in my head. Otherwise, I'll just, I'll just like babble. I need a framework. But so long as I have a framework, I feel confident when I'm standing up there and I'll just deliver it. But I understand that because there's loads of founders that especially academics, especially like really mathematical people or really deep tech people, they're not used to that.

12:51And it's difficult then to get up and be passionate about the thing that you're talking and put that across.

12:58Amardeep Parmar:And like even the other parts of the year or you said you've got, everything comes up to you eventually, right? Have you found that? How do you think your style has changed over the years? Yeah, people management is difficult. I've got to say I was quite lenient to begin with. I was quite open and I'm all about transparency and all that kind of stuff. And then you realize that actually that's a double-edged sword. It's great to be transparent, but you've got to be really smart about what you reveal to certain people. Because certain people, for the things that I'm like, other people will panic about that situation.

13:34So you've got to be sensitive to, is the right word? Empathetic? to what they might feel respond to a certain situation. That was my biggest lesson. You've got to learn how someone else might respond to something. That's quite hard to predict. I think I was too straight. I think I was too transparent before. Is that right? Too transparent, too straight is probably the one thing.

13:55Amardeep Parmar:I think it's that thing of where, as a founder, you've got to be able to believe you're just going to get through it. With somebody working for you, they don't have that same control. It's out of their control. They've got to believe that you can get through it. Yeah, exactly. And sometimes the more they know. Yeah, exactly. And if you reveal, it's a weird one because you're often told that you should reveal how you feel about stuff. I actually think, yes, that's true. Okay, fine. Be open about that stuff. But if you reveal your fears, if you reveal that something you're actually really worried about to someone in your team who is not used to that kind of stress.

14:30Amardeep Parmar:Yeah. That has a really deleterious effect. and what's worse is that you know it happens in small teams people talk and then you get Chinese whispers and it gets out of control sometimes it's better not to say anything or just say this is the situation, I'm handling it and it's quite a hard thing to do because sometimes that makes you feel quite alone on that journey and I think that's the number one thing for CEOs is the loneliness co-founders helps, I've got to say you can share that with the co-founders I have a lot of respect for entrepreneurs a lot of respect because they don't have anyone really to tend to Hello, hello I hope you've enjoyed the show so far I'm Amiri Parma, co-founder of BayHQ and the host of this show For those of you who don't know BayHQ is the community for high growth Asian heritage founders and investors in the UK Over 7 ,500 people have attended our events 200 people have been through our impact programmes and obviously there's been over 250 episodes of this podcast.

15:37Amardeep Parmar:If you want to join us and take part in the programs and come to the events, go to behihq.com forward slash join. We also just released our first ever book called Startups for Outsiders, which you can get on Amazon now and the link is in the bio. Hope to see you soon at our event. Hope you enjoy the rest of the episode. And obviously you had Sarah who joined as a co-founder as well, right? Yeah, exactly. There's three of us in total. And, you know, she's got her head screwed on, has her head screwed on. And we, you know, we have our founders meeting this Friday. And we have that on a regular basis where we just go out, we have a drink, we talk about issues.

16:16We might argue a few times, that's fine. But we unburdened. I think that's really important. It's just unburdening.

16:23Amardeep Parmar:Yeah. It's the same thing with your co-founder, that I'm in the public face with a lot of what we do. But then actual, a lot of the work and like the partnerships and the documents and all that stuff is all through him. Him, yeah. So it's kind of the interesting dynamic of where people might find out about us because of me, but then he's the one actually doing the details. Yeah. And it's like that constant alignment that we need to keep doing. Yeah, but that's a really good... I think the best co-founder partnerships are the ones where you actually complement each other's things, right? I'm terrible.

16:51I'm like, I'm terrible with detail on things like documents. Sarah's doing my head in when I don't read... I sign something in my ass, man. Sign her... You're like, well, you didn't know it's a positive way to make it. She does all the things really well that I don't do well, right? And vice versa. And similarly with chat. So I think it's important when you're finding a co-founder that you have that cadence. But communication is key, right? So again, we have a founder meeting every Thursday morning as well. It's just half an hour for us to just, not just unload, but just to say, look, this is the situation.

17:26This is what we're doing. communication is one of the most important things that early stage companies that I didn't appreciate was you know I've got everything up here if I'm not communicating that then that's it's a problem right yeah similar with everybody else again that's another skill to learn and you said at the

17:46Amardeep Parmar:moment as well because now as like things are getting further you're now kind of more focused on like you need to get the word out about what you're building because if people don't know what building, then how, one, are you going to get investment, but also customers? And like you said, okay, you now pitch more confidently or you need to force yourself to do it. Yeah. How have you been thinking about that? Like how did you get the word out initially? Like how did you get your first customers? How did you get the revenue building up to where it is? Is it more word of mouth? Yeah. So it was close network is where it started.

18:13And it's been a bit of a journey figuring out how to get revenue. Got good revenue through network, that's okay. But Now we just have to be out there and people need to see you for what you're doing and get them to be able to use your product and then have confidence in it. That's more about marketing than anything else for us. But how we actually make that work is that we're not direct B2B, we're indirect B2B through partnerships, channel partners. That way we found actually much better than direct. And again, I don't think it's very easy to figure that out from the beginning. You've got to kind of meander your way there and find that, oh, look, this works.

18:56So that's how we kind of got there. We didn't show off enough. I think that's a British problem. It might be an Asian problem as well, maybe. That we just kind of like... Double weakness.

19:07Amardeep Parmar:Yeah, yeah, yeah. It's like, okay, look, this is what we're doing. You're really straightforward about it. But when I talk to other people about it, they kind of look at this and say, Man, that's amazing. And we're like, is that? But I think we lack that Americanism. Yeah. We can do this. So often I miss your kid, but it's like, so often you see people pitching, and it's like, it's all about the product, product, but the team slide is like 10 seconds. That team slide is a bit that matters. Yeah. If you're not going into the team, and you're not telling us why you're the right person for this company, you're not going to get investment.

19:43My last chair, and my current chair also, Paul Codin and T. Kotak, they both have told me that I don't like enough. And I'll tell you now, I didn't tell you in my introduction, I literally wrote the book on AI agents in 2011 and published it. I didn't say that before. I don't show off enough about it. Because it just feels a bit dicky, right?

20:13Amardeep Parmar:The problem is that people don't say it. The ones that do say other stuff, they're They're the ones that get stuff. Yeah, yeah, exactly. Exactly. There's a balance, right? Stay, try to stay grounded. You know, one of our core company values is to be humble. Keep your feet on the ground. No one is better than anyone else. But since there is a balance, sometimes you do have to show off a bit to be able to get the things that you need. Yeah. I think it's the whole thing about is the means to an end, right? Is that the people are trying to show off because that's the end for them. They want people to think they're amazing.

20:43Amardeep Parmar:Whereas the people who think it's a means to an end, it's like, But by somebody knowing you're an AI agent, now they're going to know that you have a really deep understanding compared to everybody on LinkedIn now who's apparently an AI agent. Yeah, all of a sudden, like that. And the problem is that some of those people are speaking so loudly about it, and then people are like, oh, I need to learn from that person, rather than the person who actually wrote a book on it 15 years ago now, right? And that's the problem. There's a miscommunication, misalignment in the industry a lot where there are all these people who just speak really loudly constantly yeah and it seems for example like this right it's like with the podcast we don't generally get people on who ask to be on we're kind of selecting people yeah because by me kind of having tech are not argue but like kind of get people on yeah they're more likely to be like have the right mindset yeah whereas people who are like desperate to come on so do they just want to come on to show off or why do they want to come on right so it's always that thing about like how to frame that correctly because a lot of people listening right?