The following interview is a conversation we had with Meryem Arik, CEO and Co-Founder of Titan ML, on our podcast Category Visionaries. You can view the full episode here: $2.8 Million Raised to Revolutionize LLM Deployment.

Meryem Arik
Yeah, it’s great to be here. 

Brett
Thanks having me on, not a problem. So I see that you were a rugby player, so I’d love to start there. Tell us about your time being a rugby player. 

Meryem Arik
Yeah, it’s a pretty unusual sports to play, and I guess especially an unusual sport for a wound play. I started playing during university and part of the reason I started playing was because I love sports. Like, I’ve just always been quite an active person. But I’m also, for anyone that’s ever met me, like, reasonably tall. I’m like almost six foot and really strong. So rugby really suited me and I went into it basically because I thought I’d be decent at it. But I came out of it with fantastic lifelong friends and it’s really shaped the way I think about teams. So, yeah, it’s an unusual kind of origin story, but definitely something that I still love to this day. And I still watch a lot of rugby. Even now. 

Brett
I’ve never seen a match live. Is it as, like hardcore and as intense as I imagine? Like, how would you think it compares to, like, american football, for example? 

Meryem Arik
Zachary well, it’s like american football if you didn’t take breaks all the time. So it’s in terms of, like, the way the game is structured, it’s much more free flowing. So you get like 80 minutes of action, kind of constant action, and you also don’t have the padding you guys have in american football. So it’s kind of body on body hits, which makes it kind of raw and kind of real, constantly load of injuries. But it’s a very fun game. If you’re in Europe or Australia or anything, you should definitely go and watch it. 

Brett
Nice. Yeah, I’ll definitely do that. I watched a documentary on the. I think it was the. What’s the team in New Zealand, the All Blacks, something like that. And it was fascinating. It was very good. 

Meryem Arik
Yeah, All Blacks, the creme de la creme of rugby. But the spirit of rugby goes through no matter what skill level you’re at. So this notion of, like, teamwork and working together and no job is too small is something that every single team buys by, even the All Blacks. 

Brett
That’s awesome. And talk to us about your time at Barclays. So you were an investment banker. What did you learn from that experience, and what was that experience like? 

Meryem Arik
Yeah, so my undergraduate was in theoretical physics and philosophy. And when I was 21, 22, I really didn’t know what I wanted to do with my life. I think it’s quite hard for a 22 year old to know what to do, because you don’t know anything. One of my friends had done some work at Barclays, and I applied for a job there, not knowing really anything about banking. But the reason I decided to take the job and work there was because I really loved my boss. I really loved the guy that interviewed me, and he was just a fantastic first boss. So I learned a huge amount there. I was a rates derivative structure, which isn’t the simplest job in the world. 

Meryem Arik
I learned a huge amount during that time, but pretty much the entire reason why I decided to work with that was because of this one boss and this team. And I had a really fantastic time while I was there. When I left, I was still really enjoying it. But the reason I left was essentially, I wanted to do things that I felt hadn’t been done before. And the nature of banking is it’s quite deal flow based. So the things you’re doing one year from the next actually tend to be fairly similar. And change doesn’t happen very often. And I wanted to move to an environment where I felt like my work could have an impact that was bigger than what I was just doing at that time. I really like the notion of having compounding gains. 

Meryem Arik
So the things you do on day one will really change what you do on day ten, rather than it just being like every day you put in your shift. So that’s why I moved to the startup world eventually. 

Brett
And as you made that shift, did you always know that you were gonna become a Founder someday, or is your plan to really build a career and work in startups? 

Meryem Arik
So, my father is a Founder, so I’ve always been in an entrepreneurial setting. Like, all of our conversations at the dinner table with both mom and my dad, who worked in the tech industry, were about companies and building things. So I was exposed to it from a really young age, and I always found it quite appealing. However, some people, when they graduate uni, they’re like, I want to be a Founder no matter what. And I didn’t have it in that sense because I didn’t want to start something without feeling like I had a really good idea and I had validated something that I think could be really meaningful. And I think part of the reason why I had that requirement on myself of like, this can’t be starting a company for company’s sake. 

Meryem Arik
It needs to be something meaningful, is because I saw from what my parents went through how difficult it is to build your own business. So that’s why I always loved the idea of building something really meaningful in the world, but I didn’t want to do it if I didn’t feel like I had a really good idea. 

Brett
That’s a perfect segue to talk about what you’re building in the world. So looking at Titan ML, how would you describe the problem that you solve? 

Meryem Arik
The problem that we’re solving is that ML engineers are spending way too much of their time on building infrastructure rather than solving the problems that are really core to their business. So if we think that AI is going to be as widely adopted as we think it is, then we have a lot of work to do, and we can’t have ML engineers at every single business and every single enterprise building the same infrastructure over and over again. What they really should be focusing on is solving the problems that are unique to their business, making sure the data is in the form they need, and making sure the application is what they need for their business specific application. 

Meryem Arik
Our thesis is that infrastructure should be the kind of thing that you can take for granted and isn’t the kind of thing that ML engineers should be spending endless amounts of times working on. So that’s why we built Titan ML. We had a very strong intuition when we started it a couple of years ago that deep learning would be huge, which I think is turning out to be broadly correct. And that inference would be a really key part of the problem. And the bottleneck for solving that was building infrastructure. That would make the building and deploying an inference of deep learning models significantly easier. And that’s what we’ve been doing with the takeoff inference server. 

Brett
What gave you confidence in that intuition? Because I feel like today, as we’re speaking, it seems very obvious in this like, post chat GPT world that this was going to be a big thing. But I think in 2021. It wasn’t as obvious, was it? 

Meryem Arik
It definitely was less obvious. Back then were working with computer vision models and were working with older style transformer models, like but springs. And back then there were definitely use cases of where people are using these. So, for example, the financial services were using Bert style models a huge amount. We also of saw computer vision models being used in manufacturing and transportation and other industries. So at the time, there was a market for what were building. It was a much smaller market than what we have now. And we also had the intuition that because of what we could see from GPT-2 and then I think GPT-3 was out as well, that these generative models would also be a huge deal and would be able to solve a huge number of business problems. I think the thing that surprised us is how quickly that came. 

Meryem Arik
The things that we expected to see within about five years came within about a year, which is fantastic. But the intuition was less obvious then than it is now. But there were still early signs of the deep learning fields doing some very impressive things. 

Brett
What other founders have told me, who are in a similar space, they said that you kind of pre chat GPT, it was a battle to try to create demand and get demand going. And then after chat CPT really kind of came out into the public. It became a war and a battle to try to capture all of that demand. Is that something that you’ve experienced as well? 

Meryem Arik
Yeah, we’ve definitely experienced that. Free chat GPT. We had investors telling us that they didn’t think that NLP and language AI would be a big enough market. That was from the investor side, who are meant to have a lot of forethought, and then from the business side as well. A lot of businesses just weren’t investing in deep learning like they are now post chat GPC. I mean, it was just really a frenzy. I think what we found in the beginning half of 2023 is that businesses wanted to, quote unquote, do AI, but hadn’t figured out how to do it. So there was just like a bit of a, almost a panic. And now we’re starting to see businesses have methodologies and ways that they can prove AI value, and now we’re just working on kind of capturing that demand. 

Meryem Arik
So, yeah, we definitely see it have seen a similar pattern as those other founders. 

Brett
What are some of the tactics that you’re deploying and strategies that you’re using to capture demand? 

Meryem Arik
So we’re still relatively early, actually, most of our customer base have come through inbounds and most of them have actually come through inbound who have found me personally on LinkedIn or on kind of talks or conferences that I’ve been to. We’re in the early stages of growing out our growth function. So going from a maybe less scalable inbound model, people just come to us and recognize they see the problem with us to building a more scalable and repeatable go to market engine. So that’s one of our key focuses of Q one this year. 

Brett
How would you describe the marketing philosophy and the go to market philosophy as you make those plans? 

Meryem Arik
It’s a really fantastic question. So we are selling to very technical teams. We’re selling to teams that want to be building with open source language models like llama and mixture health. So these are technical teams and the marketing strategies and sales strategies and BDR strategies that work in general. I guess for potentially some other verticals don’t work with these developer crowds. What we have found has worked really well with developers is less salesy and more educational piece. So we are very happy to talk to our clients about open source alternatives to what we’re doing or how they can build it themselves or the underlying technology underneath it. Because if they find that content useful, then we gain trust and we gain credibility. And that’s kind of how we focus on marketing and positioning. 

Meryem Arik
It’s all about being helpful to the person that we’re trying to sell to and building credibility with them. And that’s our focus more than kind of raw, brute salesy things that we might see in other industries. 

Brett
This show is brought to you by Front Lines Media podcast production studio that helps B2B founders launch, manage and grow their own podcast. Now, if you’re a Founder, you may be thinking, I don’t have time to host a podcast. I’ve got a company to build. Well, that’s exactly what we built our service to do. You show up and host and we handle literally everything else. To set up a call to discuss launching your own podcast, visit frontlines.io Slash podcast. Now back today’s episode. When it comes to your market category, is the market category LLM deployment? Or what’s that market category today? Or what’s it maybe going to be in the future as you bring this to market? 

Meryem Arik
So the category is a really interesting question. And one of the challenges of building our particular business is that the category doesn’t really exist yet. So we struggle with how to categorize ourselves as well. So there’s a couple kind of categories that it can fit in. So one is this deployment category one is like around inference. I’ve seen in some kind of market maps an inference category, another category for serving, but none of these really seem to capture all of it because we also help with development as well. So I think what the category will end up looking like is just something fairly broad called inference solutions. 

Meryem Arik
But the category is really still to be defined, and we’re part of setting up that niche, and there’s only a couple players playing in that niche, and we’re all working to define what that category looks like. 

Brett
I know you’re still relatively early on in the journey, but I’m sure you have a specific point that you could say has been very pivotal and really changed the direction of the company. So if I had to ask that question, what’s the most pivotal decision you’ve made so far? What is it? 

Meryem Arik
That’s a really phenomenal question. So we’ve been through a couple iterations, and those have been iterations in response to the market. We’ve always had the end goal of making inference easier and cheaper through infrastructure. But the way that we’ve done that has changed over time, partly because the industry has changed. We’ve gone from looking at Bert style models to looking at much larger llms. I would say one mistake that we made early on is we got a lot of interest from a very large financial institution who really liked our solution. And there was a point which were selling and working with them. We realized that their use case was quite unique to them, and that was a huge contract that I guess would have left on the table. 

Meryem Arik
But we did a lot of reflection and decided that weren’t going to go down that avenue of building what they wanted, but actually were going to go down the avenue of building what we thought the wider problem in the market was. And that realization of, like, you shouldn’t be well hunting, we should really just be trying to solve genuine problems was a difficult process because obviously turning down very large contracts is really difficult. So, yeah, that process of realizing that what one particular well wants you to do isn’t necessarily what you should do is an interesting one. 

Brett
We’re depressingly already almost through January, so we’re recording this January 25. So I’m sure you’re already through with your 2024 planning. As we look ahead for 2024, what would you say are the top objectives that you’re trying to achieve and what’s keeping you up at night? 

Meryem Arik
So we’ve released our product at the end of Q 323, signed first deals in Q four. Now, what we’re looking at doing is scaling our go to market. That is the most important thing that I’m working on at the moment and the team is working on. So by the end of this year, we have obviously certain metrics and targets that we want to hit, but what we want to feel like we’ve gotten to is a highly repeatable go to market where to some extent not dependent on the Founder anymore. So I’m heavily involved in every single sales cycle. 

Meryem Arik
So my co-founders and I would like to get to the point by the end of the year where we’ve defined the sales process well enough that in theory, although probably not in reality, we could take a step away from that and really grow that much more. From a technical point of view and from a product roadmap point of view, we’re really excited this year to be expanding the takeoff inference server beyond just llms and towards more generative AI modalities. So looking at things like audio and audio models and vision models are things that we’re really looking forward to, that our clients have been asking us for. So, yeah, but between that and a couple other things in the works, we’re going to have a very, very busy year. 

Brett
As I mentioned in the intro, you’ve raised 2.8 million to date. What have you learned about fundraising throughout this journey? 

Meryem Arik
When we first started fundraising, we didn’t know how to fundraise. So all of us are first time founders. So went through a big process of learning and iteration along the way. I guess what I learned from fundraising was that fundraising in each round is almost like trying to find product market fit, but on like a micro scale. So you’re learning constantly how to tell your story and how to really get people to buy into what you believe in. And also you’re learning along the way from other people’s experiences and opinions and views of the market. So we saw this as an iterative process of us constantly improving our product. And in this case, our product was how were telling the story about our company and were constantly improving that. 

Meryem Arik
So every single VC meeting that I had when I’m fundraising is, I would go to that meeting and I would write down every single question that I was asked in that meeting and I would, at the end of the meeting, reflect on why did they ask me that question and what could I have done differently. So they don’t ask me that question next time. And so I quell their concerns or fears, et cetera, and I would rework my pitch. And then once you do that 10, 15, 20 times, by the end, you end up getting questions that are really insightful and really thoughtful, and that really the questions you want to ask. 

Meryem Arik
But the only reason why you’ve been able to get to that successful competitive outcome is because you’ve at the beginning iterated and learned along the way, and you haven’t been stubborn and you’ve been very agile with your methodology. So, yeah, the way that I treat fundraising is kind of how we treat finding product market fit. It’s just a constant, agile process. 

Brett
One of the things that I think has been in the news for, I don’t know, to make up a number, but maybe like five years or so has been awareness about the fact there’s just not enough female founders and women founders in tech. And that’s been talked about. There’s been a lot of noise about it, and I don’t know what the numbers are for 2023, but I remember looking at the numbers from 2022 and it was something that was just like absurd. It was, I think, 1.8% or something like that. Venture backed founders are female. What do you think can be done to address this issue and solve it? Because if you just look at the numbers, it doesn’t seem like all that noise and all of that awareness has really done anything to solve the problem. 

Meryem Arik
It’s a fantastic question and it’s something that I have thought about quite a bit. So I think the problem, there are issues in the venture backed community and there are issues in the tech startup community. However, I think most of the issues actually come much earlier. I think most of the issues that we see with women in tech comes from issues of what we see in women in STEM a university, and that comes even earlier from women that take subjects like math and physics seriously in high school. So I think almost all of these problems come from the education sector. If we aren’t able to produce the number of physics and math and computer science grads that are female in the same proportion, then we are never going to be able to create that number going into tech. 

Meryem Arik
So I guess that my thinking has been that the problem, the biggest problem, is actually much earlier on when girls are in school. I would also say, potentially something else, that being a Founder can be really hard. Well, no, it is really hard, and building a company is really hard, and nobody owes you any favors along the way. So I don’t like these schemes that feel like charity of. I feel like, oh, well, we need to be investing in women because women know this is a business and BC is a business, and we should be investing in women because they build really good companies and they build companies in ways that are different to men and have their own unique and strategic challenges. So when we think about gender diversity, I don’t think we should be thinking about it in terms of poor women. 

Meryem Arik
We’ve left them out. I think we should be thinking about it in terms of, damn, we’ve really missed an opportunity. We could have built a better business and we could have made more money. And I think only when you start framing it in terms of it making business sense will you see genuine change from the VC community and genuine change from the industry and all that. So I guess to summarize my point, I think the biggest issue is actually at school level. And the second issue is we need to be framing this not as a women problem, but as a missed opportunity problem. 

Brett
I love that answer and love that framing and I think that’s so useful and well thought out. Now, let’s imagine that I came to you and I said, hey, I want to start a startup, a similar space. Let’s say it’s not a competing product, it’s in a similar space. And I want to bring this to market. Based on everything you’ve learned so far in this journey, what would be the number one piece of advice that you’d give me before I try to bring this technology to market? 

Meryem Arik
I’m sure you get this answer all of the time, and the reason I’m going to guess you get this answer all of the time is because I got told it, which is sell earlier, build lesson, sell earlier. And I think that’s something that I think a lot of technical founders are told and technical teams are told. And for some reason, technically, teams never listen and they just want to build things. And it’s a mistake that we made before we raise money is we would build too much. We’ve now gotten very disciplined of we don’t build really much at all, if anything, unless we have tried to sell it and seeing the reaction we want to be seeing. So build less and sell more, because building is easy and finding things that people want to buy is hard and. 

Brett
Solve the hard problem first, final question for you before we wrap up here, let’s zoom out three to five years into the future, what’s the big picture vision that you’re building? 

Meryem Arik
We would like Titan ML infrastructure and Titan ML technology to be a core and probably invisible part of almost every single LLM deployment and generative AI deployment. We want it to be analogous to deploying a database when you don’t even really think about it, and it’s just kind of part of your stack, and we want to be that core part of the stack for the gen-ai infrastructure. 

Brett
Amazing. I love the vision, and I’ve really loved this conversation, and you can definitely consider me a fan. We are up on time, so we’re going to have to wrap here. Before we do. If there’s any founders that are listening in, they feel inspired. They like your approach, and they just want to follow along. Where should they go? 

Meryem Arik
So go to my LinkedIn. I post there fairly frequently, some of them insufferable posts and some of them fairly educational. But yeah, find me on LinkedIn. 

Brett
Amazing. Meryem, thank you so much for taking the time to chat. Really appreciate it. 

Meryem Arik
Thank you so much, Brett. 

Brett
No problem. Keep in touch. This episode of Category Visionaries is brought to you by Front Lines Media, Silicon Valley’s leading podcast production studio. If you’re a B2B Founder looking for help launching and growing your own podcast, visit frontlines.io podcast. And for the latest episode, search for Category Visionaries on your podcast platform of choice. Thanks for listening, and we’ll catch you on the next episode.