Breaking the AI Bottleneck: How Daniel Lenton is Powering the Next Wave of AI Startups with Unify

Rui: If you’re looking for stories, strategies, and actionable advice on how entrepreneurial careers start, you are in the red place. I’m your host Rui, and this is the start of Journey podcast, the show where every week I sit down with different entrepreneurs, experts, and thought leaders to dig deep into what it takes to get a startup off the ground.

Today I’m joined by Daniel Lanin, the founder, and CEO of Unify. Daniel has a. PhD in deep learning than spatial perception for robotic manipulation from Imperial College London consistently rated in the top 10 universities in the world and is known for his passion for diving deep into engineering challenges.

He went through Y Combinators program, recently raised 8 million in funding for Unify, which is a cutting edge start of building neural routers to send each prompt to the best large language model available. Daniel, thank you for joining me today. How are you? Yes, very good. Thank you. Busy day, but yeah, really great to be here.

Thanks for the invite. Thank you for being here. So I actually can’t believe I got your PhD title, right? It’s quite a mouthful, I would say.

Daniel: It’s quite a mouthful.

Rui: Anyway, our purpose today, as you know, is to bring your insights and lessons to our community of entrepreneurs so that they can apply them in their own journeys.

Yeah. So, without further ado, how did the idea for Unify come up?

Daniel: Good question. Well, actually, interestingly enough, we were doing something quite different originally. So we were building a project called Ivy, this, uh, unified machine learning framework in Python. And basically the idea came from just having many customer conversations and realizing that.

The world of AI has obviously shifted. Um, more and more people are building on top of these large language models instead of training things from scratch. And we realized a lot of people were just finding it very difficult to get started. There’s a lot of different models, a lot of different providers.

People weren’t sure whether to be using GPT four. Or Anthropic, or Gemini, or Lama, et cetera, weren’t sure which provider to use and really wanted to make sure they could not only get very good quality, but get things that were responsive enough and cheap enough and it became clear, I suppose. I’m not sure exactly where the idea came from, but.

Basically it became clear that there’s a huge abundance of these different models and providers, and obviously not every single prompt needs the very best model. Some prompts are appropriate for a smaller model, some are appropriate for a larger model. And actually, you know, sending these to the most appropriate model, um, on a case by case basis seemed to be something that could solve the problems that a lot of the customers were talking to us too.

So, um, it was mainly driven by, you know, identifying a problem and, and then trying to think about ways to elegantly solve it basically.

Rui: Okay, so already I see a bunch of things there that are quite interesting, like listening to those customers, seeing where the market is heading, and actually with your background and profile as well, it’s no wonder that you were able to raise 8 million.

I’m not taking any merit out because I know that, especially today it’s quite, uh, hard to do so. But I see a lot of interesting things that we can, uh, tap into in this, in this conversation. So. What were you doing before and when did you decide to take the leap? Because you mentioned this was a different project, but that was not all you were doing, right?

When did you decide that, okay, I’m gonna go all in, this is what I wanna do?

Daniel: Yeah, so with regards to the company Pivot, I mean, I guess it was two, two answers to that question. First of all, in terms of whether to dive into the company in the first place, which was two years ago, that was basically, I was working on this open source project and I just, I guess partly I just was very passionate about that and I knew that.

I guess, I mean, I’ve been doing robotics research and I’ve been doing these various projects, but I, I just fell in love with the infrastructure and I just knew that there was a lot of problems just for getting started quickly and easily, and I just really wanted to, to dive in and help to address a lot of the infrastructure challenges and a lot of fragmentation.

So I. I just, I, I just knew that it was a problem I was very passionate about and, and I kind of have always been motivated by working on something that I think is important and that I’m excited to work on. And that’s way more important to me than the funding we’ve raised or, or anything like this. Like as long as you’re happy every day and excited, that’s kind of really the objective want to, to optimize for.

So that was just quite, quite clear, quite quickly in terms of the, the pivot. So when we kind of change the company’s direction towards the beginning of this year, I think. It’s quite a long answer, but basically it became clear that the market that we were trying to address was not growing. And in fact, if anything, it was getting smaller and that the AI landscape had evolved.

I won’t get too technical with it, but basically when it became clear that the customers weren’t having the problem, we thought they were, or indeed, that they were having a year and a half before, and that problem seemed to change and and dissolve a bit. A really important thing whenever you’re doing family, a company is obviously to stay way more laser focused on the problem and not the solution.

So as soon as you start trying to shoehorn your solution, no matter how much time you’ve worked on it, then you’re doing the wrong thing. Uh, you really need to be kind of obsessed with problems and, and, you know, there was just a moment where we realized this isn’t the right direction. We need to change direction.

And, and it’s definitely been, or at least so far, the signals are that it was the right decision. So definitely happy that we made the, the tough decision. Yeah.

Rui: This is something I’ve seen happening over and over again, and I’ve seen brilliant teams, brilliant, uh, projects go to waste because there’s a certain commitment and actually resistance to change.

Right? Yeah. So I really want to dig a bit deeper into this. So when you say that the market was not evolving, that the, let’s say the need for the solution wasn’t really there, right? Not as much. Yeah. Uh, maybe not a bleeding neck, let’s say. Maybe just a pain. Yeah. What was the data that you used? To make this decision.

So was it quantitative, qualitative speaking with people? Was it the actual metrics from the response to the platform itself? So yeah. What did you use to, to say, okay, let’s, yeah, trash it. Let’s pivot, let’s go the other direction.

Daniel: I guess there’s a few things. One of which is that we’d been speaking to a lot of enterprises.

Were originally interested, I mean, basically they were using TensorFlow and they wanted to use our solution to bridge the gap between TENS Flow and PyTorch. And a lot of those enterprises that had previously been very committed to using TENS Flow and and want to use our solution to bridge the gap had then decided to migrate over to PyTorch.

So like we’d just, through customer conversations, learned that through, through ongoing relationships. Another one. Another thing with just observing the fact that PyTorch became a lot stronger. Uh, so PyTorch used to be very runtime inefficient in deployment, but then things like Torch fx, which made function tracing much better and execute torch and torch two with torch compile again, quite technical stuff here, but basically PyTorch became much more runtime efficient.

And that meant that the benefits of converting PyTorch to other efficient languages like um, or frameworks like TENS Flow and Jax brought less benefits. It used to be the case that converting PyTorch to JAX would make it, you know, five or six times faster that then reduced two times faster than maybe one and a half times faster.

So I guess like evolving. Like changes in the landscape that were there for everyone to see, as well as customer conversations kind of helped, helped make that decision. Now, that’s not to say that there’s not a market and that what we built wouldn’t have, you know, a decent number of users. But I guess like.

I’ve been very, I’ve always been very keen to solve the biggest problems and, and have the biggest impact we can. And given that we had a ton of people we’re speaking to that had a very different problem that I thought we could solve, and we were solving another problem that a much, you know, a smaller group of enterprises had.

It just felt like the right time to, to double down and solve this, this much broader problem that we knew we could solve.

Rui: Okay. So on this topic of the landscape evolving, this is something I’m seeing a lot as well, because there are, there Yeah. A lot of people building things on top of LMS actually.

Mm-hmm. And then Che GPT, uh, releases brings a new release that actually makes that obsolete. A lot of business models being challenged every day. And

Daniel: yeah.

Rui: Uh, we’re seeing some, some interesting disruption. Cool. So when you were, uh, when it, when was clear that you needed to do this pivot, was it. Your initiative, was it a combination?

The team that you had with you back in the day, uh, was aligned with you? Did you have to convince them? So what was the dynamic there to implement after realizing that this was something that needed to change?

Daniel: Yeah, I, I think it was somewhat collective. I think it didn’t require too much convincing, to be honest.

It mainly came from me. We’ve been quite lucky that our investors, for example, have been quite. Hands off and hands on when requested. So we certainly don’t have people trying to steer us, you know, stepping in to try to correct course or everything like that. People have trusted me to. Set the direction and set the vision and everything and, and so basically it was just from myself and there was no external pressure to change direction at all really.

But with the team, I think all of us got a sense that what we’re building wasn’t quite in the right direction and that it wasn’t quite getting the traction that we thought. And obviously all of us being very. Deeply technical and understanding the space knew that a few things that had happened in the space meant that our solution was less valuable.

So I think all of us were aware of this. Now, what made it quite difficult is that we actually reduced the team. So not, not because of lack of runway, but but more just because when we changed direction, what we really needed was. You know, a much smaller, more nimble team that can iterate on ideas quickly, execute very quickly without the, you know, through no nobody’s particular fault.

Just having a slightly bigger team means there is a bit more inertia, like there’s more communication bottleneck. It’s my job to make sure everybody has something to do, and if there’s not 20 independent tasks to be done, you end up kind of making up work where it’s not really needed and and stuff. So it became clear that the team size was a bit too big for kind of where we were Post pivot.

And therefore it just ended up making us much more productive actually, to make the, to reduce the team size. But that was very difficult, obviously. I mean, I’ve, you know, everyone has been incredibly gracious in understanding and you know, I remain in good contact with everybody that was on the team, but definitely kind of, you know, letting.

15 people or, or so on the team, go maybe slightly more than that was, was, was not an easy thing to do. Um, obviously it’s one of the least fun jobs as a founder, but, you know, you’ve gotta make d difficult decisions sometimes. And, and definitely for the company and indeed for everybody that that left, it’s been the right decision.

’cause you don’t wanna be part of a company where you’re not needed and not useful. So everyone’s now gone on to do, you know, great other things and, and stuff. Yeah, definitely, definitely not the easiest thing. So being, being a founder seems very glossy, but there’s, there’s, there’s difficult parts as well.

Rui: Not at all. Most of the successful founders I know, tell me that if they had known this from the start, they would never do it. Yeah, yeah. Most of, ’cause it’s, it’s mental, right? It’s just, it’s a crazy Right. And I’m, uh, I’m right there with you. I. That is for me, the worst part of the job, having to let someone go.

And mind you that I, I actually think there are great reasons to do it, not only for the sake of the company, because if it’s not successful, everyone actually loses their job. Right, exactly. So it’s better to change in the right direction than privileging. The whole and not just one individual, which is one of the hardest things that you need to do.

Yeah, yeah. Sometimes. And then for those people themselves because for sure, I mean for them it’s gonna be better for them to look for something else instead of just trying to fit a, uh, a round shape into, into a square itself. It

Daniel: benefits nobody to like be doing work that’s not strictly needed for the company.

Everybody kind of loses there, I think for sure.

Rui: Yeah. And the moment that you start making those decisions, you are closer to making better decisions and more of those decisions. Yeah. Instead of the ones that, uh, take you to the, the graveyard in of, of the startup world. Yeah. So, perfect. Daniel, you have a really cool background, right?

You were, you studied engineering and you had some robotics, uh, experience there. You were, um, at Amazon at some point in an internship, right? As a research scientist. Then you did some research on robotics as well. Mm-hmm. At a certain point you decided, okay, that academic background, that researcher background, I really want to build something, right?

Yeah. So what was, if any, your biggest fear?

Daniel: I guess? I guess it was quite a gradual transition, to be honest. So I was quite fortunate that there was never this like. Big, like now I’m gonna drop all the opportunities and try this thing that’s scary to do on my own. Like basically I’d started working on the open source repository.

Again, this was with Ivy, so the original idea of the company, I’d started working on that. Um, in parallel with my research and it was kind of a, a side hobbyist project. We opened source that got a paper out as well, and that got a decent amount of, of, of traction. And actually then investors just started to reach out on LinkedIn actually.

So we’re still in this. I think quite fortunate position where we have still never reached out to any investors. E every investor has been inbound. That’s, that’s come through or kind of mutual intros. So we never had this like ch need to kind of chase after trying to raise money or something. And, and I, yeah, and we raised the first kind of 200 k.

Off the back of the open source project and, and kind of even as I was, you know, still doing the research and in the PhD and everything. So it was, it was all quite a seamless transition actually. Yeah. So, so, so there was, it was never really that scary, I guess. Like it continues. You, you need to be a bit.

Scared in general though day to day about making sure your company’s on the right path. So I don’t think it’s any more scary at the beginning than it is now in terms of just like the responsibility I feel to the team, more to the team than myself. And also, I dunno if this is a good thing to say, but I probably feel more responsible to the team than the investors even.

Because like the, the team and myself obviously are kind of putting in our, you know, early twenties, our kind of prime time starting off our careers. And we all want to build something that’s really valuable and, and really solves a problem. And I feel. A big responsibility to them to make sure that all of their hard work is indeed building something that’s gonna amount to, you know, really being an important part of the AI stack and, and making sure that we’re on the path to do that is, is the thing that probably motivates me more than my own risk of failure as a founder or.

Feeling any sense of, obviously I do feel very grateful, you know, incredibly grateful to our investors and want to make sure we, we do a great job for them. But I think I’m kind of more motivated by making sure that our team, like is building something in, in their prime time with their engineering careers as well, to be honest.

But yeah, so you always need to have a bit of fear, but you also can’t let the fair. Drive you too much because you know what you need to be doing every day is starting the day with like excited enthusiasm to solve the problem and energy to build quickly and not like, you know, paralysis from fear of failure or something.

So you need to keep it in within balance. Obviously you need to use the nervous energy to kind of, you know, get excited and motivated.

Rui: Hundred percent. From what I am hearing here and what you said already, I think your investors would love to hear that, what you just said about the team, right? Because it seems to me that, it seems to me that you have smart money there, and that plays a huge role, right?

Having someone that can offer some guidance, not just the money, but also they don’t try to enforce their own perceptions or biased visions or whatever. Uh, right. Actually, and I can relate a lot because the line in the sand. Uh, so we do, I do everything I can for my clients. The line in the sand is if it hurts my ability to provide a safe and good environment for the team.

Right? Yeah. Yeah. So those are the most important, uh, elements in, in this equation for me. Yeah. So the environment in a startup is very different than the one you had when you were doing your internship at Amazon, right? Yeah. What are, what was. That transition? Like, was there, was it a shock? Uh, how, how did that go?

Yeah.

Daniel: Uh, I dunno, lots, lots of ways I can answer that question. I, I, I guess so obviously, I. One thing I can say about my time at Amazon and also my time in the PhD is that I learned a huge amount from my colleagues. So as, as somebody that did a mechanical engineering degree, originally my software engineering skills were, I dunno, like, were not the greatest.

Uh, I’d done like very hacked projects where it would be like one massive Python file doing some computer vision stuff or whatever. Again, ’cause my, again, my degree wasn’t software engineering, so I’d really become a good hacker, but not a good software engineer. So, yeah, so in both my lab and Amazon, I learned a ton from, from my colleagues.

I think one thing that I’ve always, and maybe this is more of a me problem. But I’ve always like struggled a little bit with not having the full freedom to kind of do exactly what I, I want to do. I felt this a little bit in the PhD and a little bit in Amazon as well. And obviously it makes sense ’cause like there’s a broader vision of like what you’re meant to be doing.

Like you are kind of a piece of this broader system and obviously like there’s a a, you know, there, there’s a lot of freedom, but there’s still like reigned into one particular area that you need to work on. And I, I, I definitely think I thrive when I am. Able to set the vision myself. I, I kind of obviously get really sometimes and it is probably just because I get very into particular ideas and I can’t let them go.

And, and, and I think sometimes that’s like pushed up against like the broader strategy of the, their vision or whatever. So I, I, I personally love the freedom of being a founder and being able to kind of. You know, make, set the vision and everything. It’s something that I feel incredibly grateful to be able to do.

I think the other thing that is obviously the case though, is that there’s just way less structure. So you need to be very disciplined and like, I just have quite a lot of conviction because I. There’s no guard rail. So when you are a founder, like it’s just you and like a lot of people that are respon like that you’re responsible for in a way.

And like you’re responsible for making sure this money doesn’t all get wasted. You, you’re responsible for making sure everyone that’s on the team is like building for something and actually like, you know, setting themselves up and everything. And. But yeah, you’re responsible for your customers as well.

And in a way, like the buck always stops with me, so sometimes it is a lot more pressure. Um, I do, I do some, the grass is always greener, right? I do sometimes romanticize about just being in like a safety cocoon of a big company where like I can just go home and have my dinner and watch the football and not think about like.

All the sales calls or whatever’s, like, whatever fire there is to put out, but you know, it on balance. I definitely like the pressure and the responsibility and the freedom and, and everything as well. I, I really thrive with all of that. I think.

Rui: The bad days are really bad. The good days are really good.

Exactly right. I mean, it’s, it’s one of those things that you’re not there cut in the middle. Definitely. So, perfect. By the way, uh, as a marketer, I’ve, I take pride in, in having built a, a few inbound engines on my own, right? Yeah. So enable. Companies to execute sales in a much easier way. Yeah, inbound investors is actually super cool.

So congratulations on that. I, I completely understand the circumstances that led to it. I’ve seen other cases, um, but it’s also something I like to really highlight because doing. Certain things, doing things in a certain way mm-hmm. Can lead to that interest. Right. Because investors are looking for money that can actually multiply, uh, their, their money.

Right. So if you can also, I’m not sure what it

Daniel: is that we did for that, to be honest, but I’m, I’m definitely very happy that that’s how the cards have fallen.

Rui: Yeah, I, this is something that most likely would have to be in a second conversation because for sure working in an open source project helped because there’s visibility on that.

But I would also like to understand how was the communication, if you guys were active on LinkedIn, if you were talking about this, if there was any buzz that you were creating. Even if it was unintentional, right? Yeah. That led to this because it’s, again, going back to the marketing fundamentals. Yeah. And I’m looking at, at investors as a potential customer, right?

Yeah, yeah, yeah. So what do they need to see in order to fill that trust, that credibility, that Yeah. Yeah. Uh, trustworthiness that will lead them to invest in the team. Yeah. But anyway, I need to go back to the, to the conversation about Unify. Yeah. So. You went through Y Combinator, right? Yes. And I would like to understand at what stage of the project this happened.

So did you have a product already? Uh, you were in the process of building it. Did it contribute to validate what you were trying to execute? Although the market, the trend of the market, I would say your idea, it’s one of the, your project is one of those where after hearing it right, I go like, yeah, that makes perfect sense, right?

Because we have. Tools like PO that have gather all the LLMs, but having a brain that actually routes the prompt to the best possible solution is, is massive. It’s is super cool. Uh, and it makes perfect sense. So. On that angle, what was the role that Y Combinator, uh, played and how did you validate the idea?

I think that’s the question.

Daniel: Yeah, sure. So, so Y Combinator was actually, so we were part of the winter 23 batch, so kind of January till March, 2023. So that was, um, pre pivot. Let’s say we really doubled down on LLMs as of basically January this year. So. It was basically a year before we did the, the final pivot.

So YC actually weren’t the very first investors. So when we did yc, we’d already raised kind of one and a half million, I think. And actually like, strangely enough, I didn’t even participate in the batch that much. ’cause we had quite a large team at that point in time. I, I joined as much as I could, but I wasn’t that present during the batch itself, which in retrospect was not necessarily the best decision because I think, I mean, I think a little bit.

I had this impression that all of like ycs advice, like didn’t fully apply to us because we were like deep tech. We were building this framework that unified all the other frameworks. It was like open source infrastructure. It’s not really the typical kind of B2C SaaS flavor that that YC really, um, you know, shine with.

But, but I, I was wrong. Basically, a lot of the, like a lot of the very basic things like don’t over hire, like really validate product market fit, don’t. Build a solution in searchable problem. I mean, in fairness to us, like the evolving LLM, the evolving AI landscape was part of it. So like we did have a, um, we were solving a real problem and like that problem kind of disappeared, but I still do think we weren’t, I.

Talking to as enough customers, like I was, I probably had spoken to like 40 or 50 customers in the whole first year and like now I’m speaking to like six or seven a day and like, I mean that’s pretty extreme, but like we weren’t taking that quite clearly enough and like we weren’t. What we do now, for example, is like every single feature we build.

Is backtracked to a customer problem? Well, not, not necessarily a feature request again. ’cause sometimes customers dunno the feature they need, but they know their problems. But it’s always backtracked to a customer problem. And that was not the case at all with Ivy. With Ivy. It was like we’re building this infrastructure with this grand vision for the future and, and the.

Know, we found the trapper become a little bit disconnected from like the actual day-to-day problems that that user base were having. Um, so you know, this is like YC one oh one. They even call it assist A solution says Problem don’t overhire. If you raise a lot of money, like you shouldn’t obsess about the wrong metrics.

Like how much money have you raised, how many employees do you have? And we never felt we were successful ’cause we had a lot of employees. But I think we had a bit of a full sense of like, we’re doing well. ’cause we’d raised a lot, we had a lot of GitHub stars like, you know. Kind of 14,000 GitHub stars and massive community and, and lots of metrics that were doing very well.

Yeah, and then I, I think kind of what I then did was like watched a lot of the YC videos, talk to the partners more. Not that they’re the only ones to give this advice, but just I think being totally honest, as a first time founder, I did fall into a few apparently very common traps. Um, it doesn’t make it feel any better that they’re very common, but I guess learning from doing is, is one way of learning, and it’s certainly the way that.

I have learned. But yeah, so I, I would say that, anyway, I’d say the YC was quite early and, and to be honest, like I think some of the, their general principles helped us make the pivot, even if it wasn’t during the batch itself or during lots of back and forth, you know, just ingesting the user guide and the videos has like, definitely.

Refined my own like intuition on these things and really helped with some of this decision making we’ve been doing recently. So yeah, only good things to say about, about yc.

Rui: Cool. Two things there that, on what you just said on the customer feedback, right. So I had the privilege of. Working with or advising or mentoring?

Uh, literally hundreds of, of founders on the, their go to market. Yeah. And this is the number one thing, speak to customers. Yeah. They actually engage with everyone, potential investors, every stakeholder from the get go while you’re building it because yeah, there’s, uh, there are things that you learn there that are gonna be paramount for success, otherwise you’ll land with something in the market and then your runway is not there for the pivots that you need to make.

So, yeah. Yeah. Start. Working on that early. Yeah. And this is the single most given advice, which is customer feedback. Speak with potential customers. Yeah. Get to know them as well as you know, your mother. That is usually the expression I use. Yeah. Yeah, because, and to your point. It’s not that you need to execute everything they say because you also, uh, risk a lot if you do everything, uh, that people tell you.

Yeah. And going back to the whole saying, people would have faster horses, right? Yeah, yeah, exactly. Not cars. If, uh, we were to do everything they say. Yeah, but getting, truly understanding their behaviors, their motivations, the pain, and what they’re actually trying to solve, not the solution for the problem itself, but what they are trying to achieve, right?

Yeah. All of that plays a huge role, not only on the product execution, but on the positioning, on the segmenting, the way that you target, uh, your addressable market and all of that. So. Thank you for, uh, for pointing that out. Then also on the vanity metrics, 100%, right? Yeah. Focusing on the wrong KPIs is so common, right?

Yeah. Yeah. And obviously there’s a direct translation here to, for, to, to marketing. I don’t care about website visits and uh, yeah. Social media followers, right? Yeah. The. Whole point is to, yeah, generate revenue as Chris Lockhead, the category pirate usually says if marketing does not translate into sales, it’s just arts and crafts.

Right? Yeah, yeah, yeah. So focusing on the right KPIs Yeah. But actually measuring the things that you are doing. Yeah, yeah. Plays a huge, uh, role there as well. Okay. Daniel, so the. First customers. So the first users, how did you, uh, acquire them?

Daniel: So it was definitely a bit of a mix. It’s, it’s interesting, I mean, without kind of going into.

Kind of all, all the granular details, like one of, one of the first customers was actually through an investor dinner. So I came to about, like all the stakeholders. It was like an eight roads dinner in London and, and like I gave an intro saying, Hey, I’m Dan, we’re doing this, we’re doing a router. And one of the other founders at the dinner.

In fact, that led to two of our first customers that that dinner. So in person events can help as well. Definitely when you’re building these relationships. Outbound definitely we, we have a outbound funnel that I do through my own LinkedIn, so I message people on LinkedIn and we use Sales Navigator to kind of find people that seem to fit, fit their right profile.

And I just, you know, send a message to hop on a call, and that’s led to. Two or three of the first. And also, um, I think the other thing is that launching quickly and often is very important. We did a kind of slightly bigger launch at the end of May, and that definitely generated a lot of inbound. So before that launch, we hadn’t had any inbound.

And after that launch, we’ve probably had like 50 or 60 companies book a call through our website and now. I dunno, kind of half a dozen of them are now using, using us. So, um, that was kind of inbound just through the launch and now we have, we have someone that started using us last week that saw a post about us on Instagram or something.

I have no idea. So I dunno, like obviously like, I guess what you want to do is, yeah, so, but long story short, it’s a mix. And obviously then the other thing I would say is that like, the approach we’ve really taken on board as well is like. Your first 10 customers that love what you do, you should really know who they are and you should really like, have a very white glove approach and, you know, iterate with them directly.

So I think I’ve, our ten first users, I’ve had probably six or seven calls with all of them. I’m constantly in email, back and forth, and like really making sure that we are hands on, integrating it for them and making it really work. Again, it’s like the YC 1 0 1, like do things that don’t scale kind of stuff.

But it’s, it’s, it’s true. And again, I think I, I, I knew that a little bit, but. It really is the case. ’cause again, the benefit is you need to like have a really strong mental model of your customer and your ICP. You need to really understand their motivations and there’s no other way of doing that really as well, apart from just actually being there with them, you know, on the ground, getting it set up and really understanding why they’re using it, what they’re using it for.

And also because. It’s also one of the superpowers you have as a startup. The fact that you can give this like incredibly personalized service is something that big enterprises can’t do. So it’s another like thing you should really leverage for the first customers. For sure. And that’s the other, sorry, just a long monologue.

But the other thing I think is that, as you’re saying about this kind of vanity metrics and KPIs, like it’s, it’s relatively easy if you do some hacks to kind of growth hack followers or growth hack, whatever, but like. What you really want is like to actually have solved a real problem. This is the really hard thing, like it’s kind of, there are ways to do all this.

So the stuff, but the really, really hard thing that almost nobody achieves is to build something that’s actually useful to a large market. Like this is something that most startups never do before they die and like. In a way, like even if you’re away behind the scenes with nobody paying attention and you are talking to all of these customers and no one gives a crap about you and you’re not making any buzz or anything, like if you actually manage to learn that and actually get the product that matters, and then they start telling their friends and you get word of mouth spreading, then you can kind of emerge from the ashes of.

You know, not being known. Be, and that’s what you, everyone’s trying to do in the end anyway. Like, unless you build a product that’s actually useful, that will organically grow, you are like, you’re not doing the right thing. And, and sometimes the really hard work getting that in the state it needs to be, is like.

Happening without, you know, mass awareness. And that’s fine. Like you just gotta believe in the process of actually building a useful thing with your early customers. And if you do that right, things will grow and you will, you know, address the market basically

Rui: 100%. And uh, again, I’ve seen a lot of people hide behind.

Solutions are actually products or brands that don’t yet exist, right? Yeah. They launch and it’s immediately the website executing the sale and they want to actually hide all the weaknesses. Yeah. And, and it really doesn’t work because then the arguments, the trust builders, the use cases, the testimonials, nothing is there Yeah.

Right. To uh, back that story. Usually it’s more about, I usually say also that the first one to 10, as to your point. Is a very different strategy than 10 to 100. 100 to 1000, right? Yeah. After you start getting those things. Yeah. So the, and I deeply believe that at the start, it need, it needs to be founder led.

Yeah. Yeah. It needs to be you. It needs, because no one else knows as much about it as you. No one. Else has the passion to go through. Yeah. I don’t know how many nos before getting a yes and then really nurturing that. Yes. As, as if it’s something really, really precious because it’s uh, right. So, and then taking in the key elements to, uh, really fine tune the positioning, the messaging, and all of that because.

In the end of the day, this is something I deeply believe. You cannot outsource that part of the, of the work. No, no. De you can’t, the execution, you can find someone to help you do things, right? Yeah. But to do the right things, it has to be the founder. And again, I see a lot of people making that mistake and going for marketing agencies or or marketing hires very early on when in my belief it should always be the founder going there speaking to customers.

You were gonna say something on this? I think,

Daniel: well, I was just gonna basically agree and, and say that at the beginning it’s, it’s incredibly entrepreneurial. Entrepreneurial anyway, because what you’re doing at the beginning is not really sales calls, it’s kind of user discovery anyway, right? Like you’re trying to work out what to build, like, and we’re still doing that.

Like it’s kind of also, for us, it’s been like a slow transition. It’s like every call still in my mind starts out as a hundred percent user discovery. Like I have no idea if our problem is useful for these customers. And you need to be pretty technically minded and understand all or not of our product to really know also if it is a, if it is relevant for them or not.

And, and like, and you, you need to be taking. You need to be absorbing as much of this information to like set the entire roadmap of the company early on. So I just think it, yeah, as you say, it needs to be founder driven and only once you have product market fit and you have a playbook and every playbook is different, like it massively depends on your ICP and your go-to market and everything.

So until you know you have a playbook that has worked for 10 people or so and can scale, you can’t even tell the person you hire of what the hell they’re supposed to be doing. So like yeah, I totally agree with all of that

Rui: 100%. And I’ll throw another thing in there. I, I mean. As a CMO, I actually execute sales calls.

So actually some of the clients we have, yeah, I was the one going there and speaking with them. Yeah. Partially because in, in current, in my current team, we don’t have, uh, SDRs or BDRs or, or any of those things. But yeah, we are fully inbound. But I go there on one hand because I have that holistic view. I can think about the solution, uh, at a high level.

Yeah. But more than that, I need to know these people again, like I know my mother so I can build better bridges between what we can do and what they need. Yeah. So. That’s something I’ll throw in as well. Yeah. So for marketing hires early on, make sure that they can sell, right? Yeah. Otherwise marketing will focus on the wrong things.

Yeah. So also, I want to ask you something. What was your approach to your point, what you were just saying. You mentioned, uh, earlier that you were doing some outbound as well, right? Yeah. And you were reaching out to people on LinkedIn. Yeah. What was the approach? Were you trying to sell the solution or inviting people to this problem that you were trying to solve?

Take me through the actual execution.

Daniel: Yeah, so, so basically we first of all get like a, a list of relevant, uh, leads through various. Ways, like one of them is like looking at the people that have liked, like Laer Index or L chain posts, people that have like LLM in their job title, et cetera, et cetera.

There’s various different, like, um, ICP filtering approaches. We’ve tried and we kind of, kind of benchmark them against each other to see which pool of, of, um, people are the most promising in terms of lead gen. But, but then, and honestly this is. Genuinely how I see the first call. Like we send, I, I send a, a personal message that’s re relevant to what they’re doing.

I look at what the person’s doing, you know, send ’em a message that’s very much tailored to them and basically just say, I would love to hop on a call and like, chat about the LM space. Basically like, you know, I think what you’re doing is really cool if, if I, if I think it’s cool, I’ll say that and then we hop on a call and it’s literally a 15 minute chat with no real agenda.

And to be honest, that’s how I see it because. I’m not sure exactly if they are gonna have a problem that’s related to what we’re building at all. So in some cases, the, the, what the call ends up achieving is it gives me an another data point for what problems people actually have or what kind of ICP we should focus on.

Maybe I, somebody’s like just not at the stage yet where they’re building L Element infrastructure and they’re just still trying to just like understand how we even apply LLMs. Then I might say, okay, well this kind of company. Maybe actually they typically don’t really have this problem. Maybe they have a totally different problem that’s an adjacent space we could like look to build into in future.

So it’s all just kind of like trying to just build a, as much of a mental model about the problems people really have and how that relates to ICPs. And then we iterate on the ICP so we can filter more appropriately. We have a better understanding of the problem space and so on. And yeah, and obviously.

Increasingly over time, more and more people as we like, refine the ICP do have a problem that resonates with what we’re doing. But even then, like the first call is literally a 50 minute touch base. Then hop on a second call a couple of weeks or a week later and that we almost always do that. And then on the second call, it’s still not a demo or anything, it’s really just like.

Understanding their problem a bit more. Now we’ve kind of, you know, made contact and, and we’re kind of acquaintances acquainted and everything. And then we start to maybe dive a bit deeper into their problem space and, and then maybe on the, on the third call, like actually demo everything. Um, and, and see if we can deploy it.

It depends though. Sometimes on the first call people are like, I’ve looked at your website. I’d love what you’re doing. I’d like to get integrated today. And of course we are then like, sure, let’s go. Um, but we don’t push it too much. Yeah. And it’s also not in our interest to do so because I think a people know when they’re being like sold to, so it’s kind of, you get an instant rejection, like, yes anyway.

And b it’s not in our interest as well. Like I, I really, um, uh, the more time they spend speaking, um, the better basically. Um, ’cause that’s the data that, you know, I know what my sales pretty sounds like. If I just give that all day, every day, like I’m not learning anything.

Rui: So, yeah. So I really wanna stress, stress this so.

This is a brilliant team. The execution, uh, from the standpoint of let’s say your, uh, typical startup project is great because what you were able to achieve in two years is great, and you are still doing things that don’t scale after two years. So this is very important because found, there’s this lure that you launch after three months, you raise money.

After three months, you should be growing, right? Yeah, yeah. Rule of 40. And there are certain things that you start talking about. Yeah. Yeah. It’s often not the case. There’s a lot of. Ups and downs going back and forth and after. If in this really trendy right problem in this Yeah. At, at the, the, let’s say surfing the wave of AI with the right team.

Yeah. With money raised in the current landscape. Yeah. You have all of this and after two years you’re still doing things that don’t scale. So yeah, taking your time to actually ensure that you set the right foundations and that you have the right elements to build a sustainable business is something that I cannot stress enough.

So thank you for that. Daniel, let’s change gears here. So are you currently fundraising? Not currently, but we might be, yeah, early next year. But let’s see. So in the previous round, what were some of the most important lessons you learned on how to deal with investors?

Daniel: Yeah, good. Good question. So our round was quite unusual in so much as it was kind of an ongoing round.

So we raised some in 2022. We then raised some in 2023 and then some more in this year as well. And actually we haven’t gone more than. Two months without raising since found in the company. So very, very unusual. I’m not saying that’s the approach everybody should take, but we’ve, we’ve basically just kind of kept relations with various investors open.

Um, I do find a lot of these conversations helpful. A because it can foster introductions to relevant other companies in their portfolio or otherwise. Also, obviously, like there’s nothing is as useful as talking to your prospective customers. This is by far the top of the list, but. I have found talking to investors, particularly those that are like Ally, talking to LLM infrastructure companies to be quite useful ’cause they do have insights or know of other relevant companies that we’re chatting to and everything.

So it is a form of staying plugged into the, the space as well to kind of touch base with, with, um, particularly relevant investors every now and then. So, I, I, to honest, I’ve never really been in pitch mode. Like, I mean, I am obviously trying to sell the startup, but. In a way I, I’ve treated every conversation with an investor as a genuine conversation where we have something to, to learn a bit as well.

And yeah, and, and I think what’s benefited us is obviously, like at the beginning we had the open source traction and everything that was um, yeah. Which we definitely was a KPI that I think fostered a lot of investor interest. But I think the kind of. Honest approach resonates. I think also people kind of same with like sales modes.

Like if you, if you treat everything as like a presentation or performance, you don’t build the relationships quite so well. So one thing I would say is like, you know, you are, you are assessing the investors as much as they’re assessing you. Like, don’t feel too much like you need to like, prove it to get their money that you, you know, don’t deserve or whatever.

Like it’s really a two-way thing and it’s just building a relationship. But yeah, I, I think the other thing we did is a rolling. Cap, which is again, quite atypical. Uh, we, we also, um, like said we don’t really want a lead that was by choice and we have quite a large number of investors, um, none of whom have invested more than kind of 500 k again, because we didn’t want to raise any, like, large tickets from one one fund.

You know, there’s different strategies here for us. I, I just have found it very beneficial to have a lot of great minds to, you know, lies with, and, and, and tap, tap into it and to brainstorm with, with different expertise and different experiences and so on. And this has really, really helped us. Obviously there’s more signatures, I dunno, in, in the grand scheme of things.

I don’t think, I don’t think these party rounds are as bad as some investors might make you think they are. And yeah. Everything for us has just been on like vanilla. Safes. Um, so relatively clean so far. We’ve not done the price round yet anyway. You asked for like, advice I’m just telling you about around.

Yeah, I, I don’t know. I, I guess just kind of in a way, obviously the product needs to speak for itself as well, so there’s no kind of tricks here. There’s no kind of, obviously if you can like. It’s true that investors typically invest more likely through a fear of missing out than like just the thing. So obviously if you can create some kind of competitive dynamics, and obviously the way to typically do that is to have a very set period that you define, this is what I’m raising, this is the deadline.

And that creates time pressure and it creates competition, which can help you know, not only the the valuation that you’re raising on, but just like the urgency with which investors act. Again, ironically we haven’t done that and we’ve still managed to raise well, I think. But I think the rolling cap for us kind of did that a bit because it was like, well, we’re only gonna raise one more ticket on this cap before we roll it up.

And that kind of created a bit of like ongoing time pressure. I dunno, these are just some, some thoughts, but to be honest, if you build a good product, the investors will come. So in a way like. The other, sorry. The final thing, I know I’m doing long answers. The other thing is like there are ways of like raising a lot of money and like people are very good at raising even when they don’t have a product to back it up.

I guess that’s something that I’ve seen some founders have, but that isn’t always in your interest anyway. Like, to be honest, like a big reason of death for companies is raising too much too soon and feeling as they need to spend all your product market fits. So I also wouldn’t stress too much about if you can raise.

5 million or 10 million or 2 million. Like if you get product market fit, you can do that with 2 million and like the investors will come if you find that. So don’t worry about it too much is my kind of general feeling, I think. Thank you.

Rui: Uh, by the way, don’t worry about long answers because I actually like to go deep into this topic, so, uh, be really at ease with it.

So then go, we’re going back to the same thing, right? Solving a real, a real problem. Yeah. Right. Having a good product and then there’s something there. About the, your, your approach that I really like, which is the, the no BS approach. Yeah. Right. Yeah. So the authenticity, uh, of it, go there, be honest, be transparent.

Yeah. Be willing to learn as well instead of putting, uh, up this facade that then will crumble at. At some point it will crumble. So it’s uh, and and it’s exhausting to, to keep it. Yeah. So go there, learn, be honest, ask what they need to see in order to put money there, because then if you execute, it makes it harder for them not to actually do it.

Yeah. And you can read as many books as you want. Cini, Kahneman. Yeah. All the persuasion that you can, yeah. Yeah. Creating scarcity, right? Those, uh, behavior and emotional triggers in the end of the day. Investors are looking for one thing, which is to make money. So show them that they can make money with you.

Yeah. And that’s it. Yeah. So again, addressing their pain point, which is, how can I multiply this? Yeah. Now that is a very different landscape than it was back in 2020 or, yeah, fascinat. Anyway, Daniel, two years ago, the first iteration of the platform. Yeah. Who did you have with you? What was the team composition?

Daniel: Uh, just me, to be honest. So when it first started, so I am not, not by choice so much as just by the way the cards have fallen. I’m, I’m a sole founder, which I guess has its pros and cons. Yeah, so it was just me at the beginning and we raised a 200 K ticket, and what I actually first did was started to bring on interns because we had this open source project and we had basically.

Well, actually thousands of totally independent tasks that could all be worked on. And we had a very clear vision of how Ivy needed to be built. Ironically, now, not what we’re working on at the moment, but it was an incredibly modular paralyzable set of tasks to do so. So bringing on interns made sense.

And then we basically had, you know, incredibly good people applying for the internships. Then, you know, some of those then got promoted to full-time roles and so on. And actually now the entire team, um, so the seven engineers on the team, everybody, uh. Apart from myself, started as an intern actually. Uh, but we now have an incredibly strong founding engineering team, so it’s quite an unusual composition.

We never did kind of like very kind of tactical, top-down hiring, but we had, I mean, we had 40,000 people apply to our role in total ’cause with this big open source. The thing, like we just had a ton of awareness, a ton of applications. We reached out to every computer science department in the world. We had this kind of big pipeline to kind of build awareness for the internship program.

So we just had exceptional engineers, you know, just recent graduates and stuff applying, and they now make up the core of the core team. So it’s been a slightly unusual hiring approach, but for us it’s definitely worked quite well. Yeah.

Rui: So what was the day in the life back then? So two years ago, how many hours were you working?

What was the focus split? How has that changed? Because some people spend 30% managing, 70% selling, whatever. Yeah. Other people building the product. How was your day? So

Daniel: back then it was obviously only myself that had built the technical project. So I spent most of my time creating documentation and videos to explain how to contribute to the project and how to expand it.

So I wanted everyone that was hired as an intern to just know exactly. Everything about the product and it just needed a ton of documentation. So I always spent a lot of time coding and a lot of time creating guides. Um, and then over time I spent more time kind of reviewing prs, making strategic decisions on the, the technical kind of setup of the projects and everything.

Over time, that has evolved and for the last year or so I’ve done, which actually I want to change, but I’ve done no coding for the last year. Basically. There’s now kind of, you know, very strong engineering team. Keep up to date with the code and check it out and everything, but I spend almost all of my effort now talking to customers.

Obviously I have a technical background, so I understand, you know, the project promise based and everything, but I spend all of my time talking to customers. Like going through the sales calls ’cause we have Firefly, so going through the notes of the customers and trying to like really set the agenda and set what the top things we should focus on are.

And then obviously doing managing. ’cause I, I then, you know, set the task to focus on everything. Um, but I, but having said all of this, I don’t think that’s perfect. I actually do think I’ve got a little bit too far from the code, not because I. Can improve the productivity. But just because I do fear over time, like there’s a risk of me losing my technical sharpness, let’s say, like, and it becomes a bit too abstracted and high level.

And you know, there’s, there’s like, I think, um, Jeff Dean, like the, you know, one of the head of AI at Google still commits to the projects every, every week. And, and I think there’s a lot of merit to that. And I’m actually noticing the team also noticing that like, ah. Just like the kind of vision stuff and like the scientific theory and everything is like totally there.

But when it comes to like the day-to-day implementational practice of it and like knowing what is a hard task to achieve and what’s an easy task and stuff like, I do think it would benefit me to be a bit close to the code. So, you know, constantly trying to tweak how my weekly composition should look like and still not got the perfect recipe yet, but, but yeah.

Rui: There’s that split is only gonna get worse. Yeah. So at some point you don’t face a, a tough decision. Yeah. But I mean, hopefully it’s gonna be for, for the right reasons. Yeah. So now how many hours were you working at the start? Oh yeah. And how many hours are you, so what’s your weekly? Uh, yeah.

Daniel: Yeah. So at the start I was working 12 hours a day, six days a week.

It’s now pretty much the same. Um, I do work Sundays, sorry, I, I take Sundays off for the last few weeks with the launch and everything. I’ve actually been working Sundays and I think it’s catching up with me, so I need to get back to having my Sundays off, I think. Yeah. So I work about, um, whatever that is kind of.

70 to 80 hours. 70 to, so

Rui: 80 hour week? Probably. Probably,

Daniel: yeah. Probably 80. I mean, I say 12 hours, but I normally, I, our morning coffee sink is 8:00 AM every day, and I’m normally working till late, late than 8:00 PM so it’s probably more like 14 hours a day, to be honest. And

Rui: there’s, there’s always an email right when you are that and you are sending it if the phone is next to you.

So, okay. So I work a lot. The, which is actually we need to be, uh, careful about it because there’s a threshold there. It’s interesting because, uh, until recently I always talked about, uh, work life symbiosis instead of work-life balance. Work life balance. Right. Yeah. If it energizes you, if it’s what you want to be doing, you shouldn’t refrain from doing it.

Sure. Then I recently had a for, uh, from another, uh, interview, uh, for, for the podcast. Yeah. I had a slightly different perspective where someone was telling me about. The risk of that becoming an addiction and the negative impact that they may, that may have as any addiction would. So for sure, there are elements in there that you need to balance, but as long as you are happy, right?

Yeah. And you are balanced. I think

Daniel: what I do think though, one thing I’ve learned is that structure’s very important. So if I. This is one of the reasons we have a morning, ’cause we’re fully remote. We have a morning coffee sink at 8:00 AM UK time every day. And as much as anything that’s to kind of, well force all of us to go to bed at a sensible time because I think what I’ve noticed is when I feel as though there is infinite time in the evening and like there’s no real deadline, like short-term deadline, then it’s a lot easier to procrastinate.

Like if I feel as though I can work till 4:00 AM if I need to, I might go and play guitar for. 20 minutes just to have a break or I might go watch a YouTube video and kind of, and you know, for 10 minutes. Then it turns into an hour and I’ve just been sat there watching YouTube instead of actually taking a break and going for drinks with my friends or whatever.

Yeah, so like. And that’s something that’s very easy to do. Obviously everybody’s aware of this, so I, I really think like, I mean now it’s like, no, at 11:00 PM every day doesn’t matter. Even if there’s like a big webinar tomorrow at 2:00 PM and I’m not prepared for it. It doesn’t matter. Like if as soon as you start making exceptions to the rule, you’re just gonna end up in this place where there’s no like set structure.

So I really just say 11:00 PM is the deadline and it can increase urgency. If it’s 8:00 PM then you’ve got a lot to do. You’re gonna get those three hours done and like you need to. Enforce that structure yourself. And we try to do that system it systematically with the company meetings and stuff. But yeah, this is just one thing I would say.

And then obviously like from 1211 till 12, I mean, that’s not that long. But then I do like, you know, definitely don’t work and you know, maybe watch a video or whatever. And also I think like taking one day off a week is, is probably a good thing. Like, like if you have one day where you totally. Unwind. So that’s also good.

’cause if you don’t, you’re gonna end up finding that downtime anyway. Like, I don’t think you can necessarily keep working a hundred hours a week consistently. Like you’re just gonna, instead of like taking your Sunday off and going for a nice hike, you’re gonna end up like watching YouTube for three hours on Sunday and then thinking, what the hell have I, why did I do this?

Rui: That’s it. The, the wasted chip. Right. Exactly. I have this opportunity, so I think you should

Daniel: accept, you can’t really do more than like 70, 80 hours. The time you have left, like actually segment it and do something like worthwhile go see somebody. Yeah,

Rui: yeah.

Daniel: Be intentional,

Rui: right? Yeah. Be intentional about it.

Yeah, exactly. Okay. Still, it’s still 80 hours a week, so Yeah. How did you balance, how do you balance this with your personal life?

Daniel: Yeah, it’s, it’s difficult. I mean, obviously the professional bleeds into the personal a bit. We’re all very close on the team and everything, so you know, we have a lot of camaraderie that has.

So there’s some personal aspects to just the team in a way, to some extent, you know, very, very good friends with everyone on the team. But, but yeah, outside of that, I think again, it’s just a case of being quite strict. So like. Taking Sundays off, making sure you make plans to see your, you know, family and friends and getting things in the calendar as well is one thing I do.

’cause I do have a bit of, I think every founder probably has some, like, let’s say workaholism tendencies. So what you need to do, I think is be quite deliberate. Make plans to go for drinks next Friday, make plans to go out for the weekends, you know, and, and then. It’s in the calendar and stick to it, which is when you’ve got, you know, deadlines and fires to put out, it can be tempting to, to just segment every hour for work.

But you gotta, like, again, it’s like a marathon, not a sprint, right? Like, it feels like you’re constantly sprinting with fires to put out. But having done this for two years now, like if you burn yourself out for like two weeks, like it’s not in the company’s interest. So you gotta like, you know, you gotta sustain yourself.

Yeah,

Rui: 100%. And yeah, the. Most successful people, some of them tend to obsess over these things and it is what it is, right? So, yeah, again, it’s a matter of keeping the right structures, as you said, and the right balance, I would say. So Daniel, we are almost at our time, so I have a couple of rapid fire questions.

Yep. So these are straight to the point. Short questions, short answers. Yeah. So the first is. What was your biggest mistake to date?

Daniel: Biggest mistake I think was not pivoting earlier and over hiring. Um, probably, yeah.

Rui: What was your best

Daniel: decision to date? Maybe pivoting again in the context of the company. I would say like having the conviction to change direction was probably the best decision so far yet.

Can you share one

Rui: key

Daniel: lesson

Rui: you learned on product?

Daniel: On product? It’s quite vague. Again, just you don’t know what you, you don’t know. Your customers know. My guess is the key lesson. Like listen to them, not your own intuition. Too much.

Rui: Yeah. Okay. One key lesson in marketing. I think it’s gonna be similar.

Daniel: Yeah. Kind of. I mean like I guess the more point is like a great product sells itself is one of the main things, to be honest.

Rui: Can you share one key lesson you learn on managing people?

Daniel: Um, I think what, what I would say is that give people autonomy and don’t micromanage. I think more people go too far in micromanaging and like, let people fail, let them make mistakes, let ’em learn from them and trust them to grow into the role.

Rui: Love it. Can you share one key lesson you learned about creating and maintaining a good culture?

Daniel: I think. Openness is very important. So one thing is we have a weekly sync where everyone on the team shares their thoughts, what we could be doing better as a company, be okay to receive criticism from your team.

As a founder, it’s not personal. Everyone wants the company to succeed. So the more you can democratize and make people feel ownership, the better.

Rui: So one resource that was invaluable to your success, and this can be a book, a podcast, a

Daniel: mentor. I mean, it’s a bit, it must be very common, but I, I do just think like the YC startup school videos have actually been like, super on point, um, more so than I actually thought they would be.

So they, they’re very good.

Rui: Perfect. So last question, I’m gonna put you on the spot. So who should I interview next? Someone you really respect?

Daniel: Oh God. Yeah, good question. I mean, there’s, there’s a lot of, there’s a lot of amazing founders. I, I think, um, Joe or Jao at Crew AI has done an incredibly good job, uh, if you managed to get, if you manage to lock him down, uh, they’re doing amazing.

Yeah. A tie, co-pilot kits doing really amazing work as well. I mean, I know people in the AI infrastructure space a lot, but May maybe those are the two that come to mind immediately. ’cause I chatted with ’em recently. Yeah.

Rui: Perfect. At some point we will speak about maybe, uh, an intro or something along those lines, if you can.

Of course. Yeah. Yeah, of course. That’s it. Thank you, Daniel for taking the time out of your schedule to sit down with me. This was really amazing. As a last note, where can people find you?

Daniel: So if you go to unified.ai, that’s where you can find us on the website. And then also on our Discord, if you go on like company at the top of the page, I’m on Discord pretty much all the time.

So you can find me there and you can also book a call through the website, et cetera. So lots of ways. Yeah. Cool.

Rui: Thank you Daniel. So to you listening, thank you for staying with us for the full episode, and I’ll see you again in the next edition of The Startup Journey podcast.