The following interview is a conversation we had with Gabriel Bayomi, CEO of Openlayer, on our podcast Category Visionaries. You can view the full episode here: $4.8 Million Raised to Power the Future of Machine Learning Testing

Gabriel Bayomi
Thank you so much for having me. 

Brett
Actually, not a problem. So, to kick things off, can we just start with a quick summary of who you are and maybe a bit more about your background? 

Gabriel Bayomi
Absolutely. I’m Gabriel. I’m a brazilian engineer. I live in the US now. And yeah, I’m the CEO, Co-Founder of Openlayer, where we work with making AI safe and sustainable. So very excited about that. 

Brett
When did you move from Brazil to the US? 

Gabriel Bayomi
Yeah, so I studied here. I did my undergrad partially in Brazil and partially in the US. So I started both in University of Brazilia and also at Cornell. And then I decided that the US had way more opportunities if you want to follow the computer science career path. So in 2017, I moved here for good. I started my grad school at Carnegie Mellon. At some point in grad school, I was like, hey, I really want to go to industry. So I got my master’s from CMU and went to work at Apple, working with machine learning and AI, more specifically, I was working on the Apple Vision Pro for a long time. I couldn’t even talk about it because a secret project and couldn’t talk much. But basically working in AI for the Apple Vision Pro, I noticed there was a gap in the industry. 

Gabriel Bayomi
Know building models was not the hard part. The hard part was everything around it. How do you test to make sure it’s good and safe? How do you monitor it in production to make sure the performance is as good as you expect, and also the different evaluation metrics and scenarios and tracking experiments. Doing all of that was way harder than actually building the models. So I talked to a few of my friends at Apple and like, hey, would you like to quit to start something to fix this problem once and for all? And they said yes, and we left to start Openlayer. 

Brett
Amazing. When it comes to inspiration, for you, who is the most inspiring Founder that’s out there for you? 

Gabriel Bayomi
Yeah, that’s a great question. So for me, I think would be enhicido Buddhist. I think you actually interviewed him at some point here. And he’s a brazilian Founder as well, went through y combinator. And he’s just, like, a very courageous person because he was an entrepreneur before, back in Brazil. Then he was like, yeah, I want to go to the big leagues. I want to go to the US to start a company as well. So he dropped out of Stanford to do it, and he started Brex. He’s the Founder of Brex. He started at the same time that I was in grad school. And it was a big inspiration for me because it was, like someone that came from the same country as I did, doing something so cool going through y combinator. 

Gabriel Bayomi
So I was like, hey, okay, I’m going through industry, but I want to always keep my eye on going through IC. So actually, when we left, Apple went directly through y combinator, which was a dream come true. And a big part from his inspiration. 

Brett
Is he like a God in Brazil. 

Gabriel Bayomi
He’s very well known. I would say. I would say he’s probably the most well known Founder. People really admire him because he started actually, a big company in Brazil, left to start even bigger in the US. So most founders really admire him. 

Brett
Yeah, makes a lot of sense. And I think there’s a lot to admire there. That was definitely a fun interview. And, yeah, he’s a fascinating entrepreneur, and everything he’s built is fascinating. And he’s also, I think, quite young, too. Right? I think he’s under 30, if I’m remembering correctly. 

Gabriel Bayomi
Yeah, I think he’s, like, 27 and 28. So very young guy, but very successful. So super cool. 

Brett
Yeah, that’s amazing. What about books? What book would you say has had the greatest impact on you, not just as a Founder, but as a person? 

Gabriel Bayomi
Yeah, that’s a good question because I would say, like, two very different answers. So, in terms of personal, there is a great author in Brazil called Mashado Jacobs. I know I’m doing this thing all about Brazil, but I promise it’s not the case. But he had this book called the Posthum Memories of Brasco Bas, where it’s like the memories of someone that already died reflecting on their own life. And it’s just a fascinating book going to every step of your own life and reflections about what he did right, what he did wrong, and everything. And that’s my favorite book by far. But in terms of my Founder journey, I would say it’s not even necessarily a book. It became a book in the end. 

Gabriel Bayomi
But I don’t know if you’re familiar with Jason Lemkins from know, this compilation of posts that he made on quora and then he created kind of a book on top of it. And this has really changed my journey as a Founder because such an interesting and in depth understanding of how SaaS business work. And I would say like, this is my go to book. I have it on the website. I also printed to have in my hand because there is not an actual print version. It’s an online book, but I want to have it anytime that I’m endowed to look and go in depth. But I don’t know, I would really encourage anyone working in software as a service to take a look. I think it’s best of SAS. I know this sounds weird because not really a book, but man, it’s amazing. 

Brett
Yeah, I follow a lot of his writing and just the content in general, and he’s so good. And there’s no better authority on SaaS than Jason. 

Gabriel Bayomi
Absolutely. 

Brett
Yeah. Let’s switch gears now and let’s dive a bit deeper into the company. So I know we touched on that there in the intro at the start of the conversation, but let’s expand on that. So when it comes to the problem, what problem does Openlayer solve? 

Gabriel Bayomi
Yeah, it’s a great question. I think the main problem that Openlayer solves is related to how hard it is to test machine learning models. Because if you’re going to test software like normal software, you know what the output is going to be and you have an expected input and then you just check if one is the other. Right. With machine learning, you have this statistical nature of it, so things can vary. It’s not guaranteed to work in a particular way. Not always. The same input will give the same output. Doesn’t mean that’s necessarily wrong. So how do you create a system to test this, both before you put something out there in production and also when something is already there in monitoring, how do you make sure that things are working correctly? And this becomes even a bigger problem. 

Gabriel Bayomi
When you’re talking about AI and llms, you want to make sure if you have a chat bot powered by AI, you want to make sure that know hallucinating answers saying like, hey, the best way for you to get started is to join this website that doesn’t exist and click on this link that doesn’t exist. This is very problematic. So our work is basically making AI safe and performant. 

Brett
Super interesting. Take us back to the early days. It looks like you founded the company in June 2021. What was it about this problem that made you say, yes, that’s it. I’m going to go and dedicate my life to building a company around this. Or at least dedicate the next five to seven years of my life to building a company around this. 

Gabriel Bayomi
Yeah, that’s a great question. So I feel like it was honestly born from frustration because it was like our jobs back at Apple to do the error analysis of these models. We really would serialize crane and keep thinking, when does it do it the right thing? When does it do the wrong thing? What are the scenarios that it works? Well, can you come up with a list of scenarios that we should always test? Can you create some kind of unit test out of like the job was super interesting, but it was also super frustrating because there’s like no framework to do when you think about code. We have GitHub, we have GitLab, we have different CI CD pipelines, we have so much things that can help make your job easier. 

Gabriel Bayomi
But when you talk with an ML engineer or data scientist, usually they’re using like a Jupyter notebook, they’re just writing out on this very bad infrastructure, trying to make something work. And out of this frustration we decided to like, you know what, let’s apply with exactly this idea to Y combinator. Let’s apply with the string. That’s the most painful in our day to day job. And we got in, decided to quit a relatively young team all in our twenty s, and we been having fun in the journey. 

Brett
Wow, that’s super cool. What did you learn from your time in Y combinator? 

Gabriel Bayomi
Yeah, so YC has this very strong mantra which is ship code and talk to users. And I think it’s so easy to try to do other things in the early stages, try to create a company culture, or try to create this particular, very intricate marketing scheme, or trying to hire these particular professionals when in the very beginning of the journey, when they’re an early stage company, the only thing you need to do, at least in technologies, is to ship code and talk to users. And I think this insistence, like insisting on this advice for early stage companies of like, hey, we don’t need to worry about culture, don’t need to worry about this or that. Of course these are important things for a company, but in the very early stages, just focus on shift code, talk to users, iterate, and keep doing that forever. 

Gabriel Bayomi
And eventually you get there. And I think as a Founder of course, we made so many mistakes as first time founders, but whenever we are unsure about what to do next, we just think about, okay, we don’t know exactly what the best next step is, but as long as we’re shipping code and talking to users, we know you’re going somewhat in the right direction. So I think this was the most useful experience in YC. So it was basically three months of this advice, which really getting to your mind. 

Brett
Let’s talk a little bit about customers. So I see on the website, ebay is the logo that really stands out. So maybe we can talk about them or just any of the paying customers that you have. What’s that process been like to land those first paying customers? Because that’s something that all founders, all startups, they really struggle with. How do you get first people to give you money? So what was your journey like there, and what did you learn along the way? 

Gabriel Bayomi
Yeah, so for our first set of customers, we really got them from investor intros. Actually not even investors. When were not even fundraising people that in the end became investors but already believed in us, made some interest. And that’s why we got our first group of paying customers. So I don’t have a specially cool story there, but I think that one thing that we learned that was interesting this early part is like, as a software engineer, if you went to a good school, you’re a software engineer and everything, sometimes people are a little bit scared to ask for help, to ask for that new intro or to ask for what they really want. And I think becoming a Founder is a very humbling experience because we learned, like, we can’t do this alone, so we need to ask for help. 

Gabriel Bayomi
So we asked people that were not even invested at the moment to be like, hey, can you introduce that other person that have no way to access? Can you get us there? And just by doing that, just by taking the courage know, asking for favors, basically in these intros, were able to get our first group of customers, including eBay. 

Brett
Wow, that’s fascinating. 

Brett
This show is brought to you by Front Lines Media, a podcast production studio that helps B2B founders launch, manage. 

Brett
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’ve 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 podcast. 

Brett
Now back today’s. 

Brett
Episode. What about the competitive landscape? What does that look like today? 

Gabriel Bayomi
Yeah, that’s a good question because, as you know, AI is changing a lot and changing very fast. I’m going to divide things into traditional ML, traditional AI, and the generative AI boom that’s been happening lately. So I would say, like, in the competitive landscape of traditional ML, there is a few companies in the Mlops space. Some companies work on experiment tracking. Some companies work on monitoring. Specifically. Some other companies work on testing of ML systems and on the generative AI space. It’s a very new one. And there’s a new group of companies which they call themselves LLM ops, basically like operations related to llms, which try to do similar things to these ML ops companies, but more focused on generative AI. 

Gabriel Bayomi
So they do things like prompt tracking, prompt testing, monitoring your tokens and your latency of your request to OpenAI or to cohere other providers. So I would say our company, were born in a traditional ML, but we noticed the change for generative AI was happening very fast, so we decided to also join this phenomenal boom. So we are like, in both sides, we both have competitors on the traditional side of AI and also on the generative AI, some companies doing the pump level operations. 

Brett
What about when it comes to marketing? What are you seeing work right now? And what are you doing to rise above all that noise that exists in this hyper competitive environment? 

Gabriel Bayomi
That’s a really good question. That’s a very hard problem. I’m not going to lie. I’m not going to say, like, hey, we solved this problem for good because we have the best messaging, because a lot of the pitches for AI, like LM ops or ML ops companies do sound a little bit similar. Like, hey, I want to make this safe. I want to make this performant. I want to make sure that you’re not going to hallucinate. So one thing that we’ve learned is to get more direct to the point of what we exactly do. Instead of going to the abstract idea space of like, we make your AI safe, we try to market things more directly. For example, hey, get alerts when your LLM fails. Like, very direct to the point. 

Gabriel Bayomi
And we noticed that it strikes more authentic with a lot of our audience. Instead of like, hey, we’re going to solve your hallucination problems. On the safety AI space, we’re actually trying to be like, hey, we can really help you do this thing here easier. So that’s something we are always learning. But one of the things that we notice is like going directly to the point of what we actually do instead of going to the abstract world of things. 

Brett
Does it feel like you’ve reached product market fit, or where do you stand on that journey? 

Gabriel Bayomi
Yeah, that’s a very interesting question about product market fit. I actually have an interesting story. So, as you know, Paul Graham was the Founder of Y Combinator. And y combinator is one of the companies that really talk a lot about product market fit and getting there and make sure that it works. So once we had office hours with Paul Graham, were like, hey, let’s talk with him and let’s learn a little bit. And we started talking about product market fit here, product market fit there. We need to get there. And he really said, I don’t really believe in this idea of product market fit. And were so shocked. It was like, hey, I thought that YC, the whole thing was about product market fit. And he said, I believe in things working and companies surviving, and that’s it. 

Gabriel Bayomi
It’s all about being default alive, not default dead. Not necessarily product market fit. That was so interesting to me because I’m not going to say that I totally agree with that. I still like the idea of product market fit, of really getting this pain that can grow fast and make things work. But I think that the definition changes so much that it’s hard to say. Right. I think we have a strong product in a growing market, but especially in the early stages, I think I would be wrong to say, hey, we have product markets fit 100%. I think we are learning a lot in a space that’s changing every month. And I think our product, clearly our users really love the product that we have built. If that’s product market return, no, you want to learn in the next few months. 

Gabriel Bayomi
But I think that’s where we stand. 

Brett
Right now, makes a whole lot of sense. As I mentioned there in the intro, you raised almost $5 billion to date. What have you learned about fundraising throughout this process and throughout this journey? 

Gabriel Bayomi
Yeah, we learned a lot about fundraising, especially through I combinator and also through other founders. What I would say is all about, I think there are three things that, at least in this seed early stage that investors really care about, which is the team, the market, and the proof. So basically you have a strong team that can execute. Is the market big enough so it makes sense to execute on this market? And what’s the proof so far? What’s the progress that you have made that shows that this team can perform in this market? And we focus our whole fundraising on showing these three things on showing like, hey, there’s a strong team out of Apple that have really dealt with this problem before several times. The market is huge and is growing even more. 

Gabriel Bayomi
And now, as you can see, the whole AI problem is like, basically every company is becoming an AI company. So it’s a huge market. And also focus on like, hey, in this very early stages, we’re able to get these early customers, big brands and everything. So I think it’s a compelling story from the beginning. 

Brett
Let’s imagine you were starting this again today from scratch. What would be the number one piece of advice that you’d give to yourself? 

Gabriel Bayomi
I would say to just ship code and talk to users. Don’t get distracted with anything else. Although we did like three months of I combinator to tell us that, I would say it’s too easy to get distracted. And especially because we started in 2021 when it was this crazy year of venture and things happening here and there, it’s easy to lose track of what really matters. Do your users love your product, especially in the early stages, more than revenue, more than anything else? It’s like, do people really love your product? And it took a while for us to figure out that and be like, hey, this is the only thing that matters now. 

Gabriel Bayomi
Later we can care about the other problems, and now we’re super happy because we have a good group of people that really love what we’re doing and we’re so excited for the next phase of our journey. 

Brett
Let’s talk about that next phase of the journey for our final question. So let’s zoom out three to five years into the future. What is that big picture vision that you’re building? 

Gabriel Bayomi
Yeah. So our vision is that any company using AI has confidence on what are the things that my model does well, what are the things that my model does not do well, and what are the remedies that I’m using to fix that in particular? So I feel like a lot of the ML and AI nowadays, people just put it out there in production and hope and pray that streams work out. And what really want to change is like, hey, I’m putting this out in the word. I know it does really well for women in this kind of household with this level of income, and it does really poorly for men between twelve to 19 years old in this other sector of society. And I know the shortcomings, and I have the guardrails in place to make sure it’s not going to happen. 

Gabriel Bayomi
So basically to become the guardrail of the AI revolution, ideally people think about OpenAI and cohere and all of these companies building these huge models and think like, okay, what’s our Openlayer stack? To be able to mitigate all the issues that are going to come naturally from deploying it. So basically, the way that people think about code and creating tests, unit tests, or CI CD pipelines, they’re going to be thinking the same on AI, but using Openlayer. 

Brett
Amazing. I love the vision, and I’ve really loved this conversation. We are up on time, so we’ll have to wrap here. Before we do, if there’s any founders that are listening in and they want to follow along with your journey, where should they go? 

Gabriel Bayomi
Yeah, I should go to our LinkedIn Openlayer and also my personal page, Gabriel Biomi. I’m always mostly reposting Openlayer stuff, so yeah, excited to connect with folks. 

Brett
Amazing. Thank you so much for taking the time to chat. Really appreciate it. And again, really enjoyed this conversation. 

Gabriel Bayomi
Thank you. 

Brett
All right, keep in touch. 

Brett
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.