Part I: From MIT Researcher to YC Entrepreneur. Building Workflow Stack with AI w/ Bernard Aceituno
Today’s guests — PhD, Bernard Aceituno
Co-Founder @ StackAI (YC W23)
The following is Part I of my conversation with Bernard Aceituno, Co-Founder of Stack AI (YC) and previously PhD at MIT. I will release Part II at another point as we will do the recording again. Here is a snippet of our conversation at MIT CSAIL, where Bernard spent 5 years researching.
Summary
In this engaging conversation, Bernard shares his eclectic journey from Venezuela to becoming a co-founder of Stack AI, detailing his academic background, entrepreneurial spirit, and the challenges faced in the startup world. He discusses the importance of collaboration, the evolution of AI in industry, and the significance of understanding customer needs. The conversation also touches on the dynamics of building a team, the role of research in product development, and the future of AI in enterprise automation.
Takeaways
Bernard's journey began in Venezuela, where he pursued his passion for science and technology.
He transitioned from academia to entrepreneurship, driven by a desire to impact the world with technology.
The importance of being in an entrepreneurial environment to stay motivated and focused.
Collaboration with Tony led to the creation of Stack AI, focusing on solving real-world problems with AI.
Y Combinator provided crucial support and validation for their startup idea.
Understanding customer needs is essential for product development and success.
The shift towards enterprise automation presents both challenges and opportunities for startups.
Building a strong team with shared values is critical for growth and success.
Transparency and explainability in AI are vital for building trust with customers.
Immigrant founders often face unique challenges but also have access to valuable mentorship opportunities.
“We both knew we wanted to leave the academic world. We took a big bet at the beginning. We realized that what made sense was having more of a workflow builder tool. The moment we released our first launch online, we got 10 customers in a week.”
“It’s very important to be in an environment that allows you to be in a mindset.”
Chapters
00:00 Introduction to Bernard's Journey
02:39From Venezuela to MIT: A Unique Path
06:00The Entrepreneurial Spirit: Challenges and Decisions
10:30The Birth of Stack AI: A Collaborative Vision
15:56Navigating Y Combinator: The Startup Accelerator Experience
20:50Understanding Customer Needs: Shifting Focus in Product Development
25:08Enterprise Automation: Challenges and Opportunities
30:23Building a Team: Key Qualities for Growth
34:24The Importance of Support in Entrepreneurship
37:37Research-Driven Products: Balancing Innovation and Market Needs
41:41Market Dynamics: The Evolving Landscape of AI
44:51Transparency and Explainability in AI
49:20Reflections on Immigrant Entrepreneurship
Transcript
Alp Uguray (00:01.622)
Hi, everyone. Welcome to Masters of Automation. Bernard, it's a pleasure to have you today with me. It's also the first in-person podcast. So it's great to have you here. Today, we're at MIT. It's actually where you and Tony met for a long time ago. But just to kick things off, you have an incredible career. mean, you started as a researcher.
Bernard (00:07.342)
Thank you.
Bernard (00:11.118)
Why should we here?
Bernard (00:20.396)
Yeah, indeed.
Alp Uguray (00:31.352)
Right. then you had before you start focusing on Stack AI, you had another startup as well. Many years before, yeah. Different lives. But tell me more about like how, where did your career start and then where did your passion lead you to to Boston?
Bernard (00:31.715)
Yeah.
Bernard (00:38.958)
Yeah, but many years before, a long time, a different life.
Bernard (00:52.481)
Yeah, so I have a very eclectic career. I'm originally from Venezuela. I was born and raised there. I grew up in a home where I was always invited to pursue my passions and my interests. And I was very interested in math, physics as a young kid, loving music. And when the time came to go to college, I couldn't leave Venezuela at the time, so I had to go to the university in Venezuela.
For a long time, I had to start my career there. I always knew I wanted to be some sort of scientist through that time. And so I studied computer science and electrical engineering at the time, which I knew would give me the breadth of opportunities to kind of like have a background in math, but also in computer science. And at the end of that journey, I threw all the political crisis in Venezuela when most of the universities were canceling semesters through protests and stuff.
I decided, okay, well, I have a lot of free time. I'm spending six months at home. I might as well start doing research. So I started doing, I kind of like learned everything on algorithms, planning algorithms, machine learning, starting since 2012. And since then I've been involved in the field. I started doing research kind of like on my own at that time. I published my first papers with the help of my undergrad advisor at the time, who was a slightly old professor. And slowly through that, I built a research career. That led me to come to MIT to do my master's degree where I, know, was...
I started doing planning algorithms for robotics, which is slowly turning into more like applied math research towards geometric problems that were very relevant to machine learning optimization. After that, I my PhD. I spent a small sting of time in Facebook research when I worked on the first type of transformers, mostly applied to reinforcement learning. And it just wows me to see how much things have evolved since then, which was 2020. It was just four years ago, until now.
Alp Uguray (02:16.792)
Good morning.
Bernard (02:40.972)
And it's not like I was in a small lab. I was in probably one of the most well-funded research labs in the country. And just with the pace of four years, it's been amazing to see how much it has evolved. Fast forward, Tony and I became very good friends at MIT since we started. And yeah, we both knew that we wanted to, at some point, leave the academic world and go and bring this technology to the outside world. We both have been fascinated by automation for a long time.
We both have been founders before. He did some work on devices for safety and writing. I did some work on fintech back in Venezuela before I came to MIT. That was a very long time ago. Essentially, I don't know how familiar you are with the economics of America, but back then, Venezuela had a specific problem, which was that it was illegal to do transactions in foreign currency. And the country is going through an hyperinflationary period.
Alp Uguray (03:15.608)
What was that about?
Bernard (03:40.384)
with high devaluation, meaning the currency lost value extremely quickly. And there was going, we had a ton of inflation at the same time. So it was impossible to live with the current foreign currency, with local currency. And it was illegal to do transactions in foreign currency. It means you can use US dollars or euros or any more stable resource. Back in 2017, this law didn't apply to cryptocurrencies. So me and my friends, know, who had been kind of interested in for a while, we said, well, it could make sense to help companies.
Alp Uguray (04:00.248)
How many users did you get?
Bernard (04:08.204)
small businesses do transactions in cryptocurrencies, at least with stable coins or with coins that could allow you to make fast transactions. So, you we started working with that with a small app and, we grew pretty quickly, actually. We had 5,000 businesses in Venezuela using us at one time and more than, you know, maybe like 100,000 people kind of like using it at that time to do transactions. We had some regulatory issues around the time I got into MIT, among other schools to come to grad school.
And I kind of had to choose, okay, do I want to keep pursuing this, which has a very low chance of success, where I'm going get in trouble with government? Where I know I'm have to leave all these issues, or do I can like go and pursue my academic career and then try again in Silicon Valley or something? So I took an idea. I always knew that I really enjoyed the idea of building something, shipping it and getting people to use it. while I was spending my time here, I kind of followed that mindset through the years.
Alp Uguray (04:42.163)
Yeah.
That's incredible. mean, when you think about it, you solve not only an opportunity at the time, but you also solve a big problem both in the country and the way people live with inflation. The making decision to pursue your career is definitely a tough choice, right? Like leaving the entrepreneurial dream behind, which also ends up being
Bernard (05:22.122)
Yeah, absolutely, absolutely.
Alp Uguray (05:27.928)
You'd be an upcoming entrepreneur anyway. Tell me more about that part.
Bernard (05:30.048)
To be honest, I always knew I wanted to be a scientist also. I was on the fence. I still have a deep passion for science. my free time, I know. I love mathematics. I love physics. I love algorithms. I read research papers on a daily basis still. Even though I might not apply for Stack AI, maybe completely unrelated to Stack AI, I'm still always trying to stay on the fence, on the loop with Stack AI. People still email me about my research. They ask questions. I'm always on my email looking to help them understand the algorithms I made.
Alp Uguray (05:56.824)
Yeah, I mean, it is interesting because it's like, a way, you did the startup and then you came to MIT to pursue your research. what was your research about?
Bernard (05:59.66)
And yeah, I miss the days where that was my full-time job. Now I'm doing something where I can see the impact in a much more immediate way.
Bernard (06:24.044)
So I do all sorts of things. you need to go into a lot of detail. I would say that I work in two specific fields. I work in planning and optimization and reinforcement learning, maybe, so I can have a subset of planning. And I work heavily on algorithms for planning and how this could be designed to have certain properties and certain geometries. And I feel if we into too much detail, we'll get into a two-hour conversation that make your audience leave the call.
Alp Uguray (06:46.273)
Yeah, totally get that.
Bernard (06:53.76)
Yeah, so I did a lot of algorithms and optimization. And then I had some applications to reinforcement learning. And when I went to Facebook, I researched that actually was kind of where I started pretty much my efforts, taking some of the algorithms I made and adapting them for the AI world.
Alp Uguray (07:07.928)
So let's talk about them. Because what is very interesting is, especially having a research background and then the law for physics, mathematics, then over time in the startup life, translating into more meetings and then customer support tasks and like more customer success type of things like that becomes like an everyday. Especially like you're transitioning from a researcher to
Bernard (07:15.436)
Yes.
Bernard (07:24.15)
Yeah.
Bernard (07:30.028)
Yeah.
Alp Uguray (07:35.368)
an entrepreneur, what were the some things that shocked you? Like that was different.
Bernard (07:40.995)
shocked me. Wow, that's a good question. I knew I was a beginning entrepreneur before and I knew that it was just eating mud all day. That's how they call it. That's how say it. I knew it was going to be a lot of hard work. I noticed a lot of doing things that don't scale at beginning. I always felt comfortable with the idea. Even on my research, I was always into doing things that don't scale approach to life. That was something that I realized. Something that shocked me as an entrepreneur is
how important it is to be in an environment where you are surrounded by other entrepreneurs or people who live like entrepreneurs. And I'll tell you why. And that's actually one of reasons why I think it's so important to be in a set for many people. When you live a life where people around you are not working 24-7, are not making little money, are not embedding their lives into a long-term project, you get demoralized very quickly. Because you don't have that vision of everybody around you, they have a job, they take the weekends off, you know, they...
They can go out for dinner every day whenever they want. They have vacations. You see, okay, why am I doing this? Am I crazy? Am I wasting my time? And that demoralizes you a lot. And the tourism startup is very hard. It's one of the hardest things you can do in life because both is very intense. It has a very delayed reward towards whatever you do. And you need to be very driven by the passion to solve the plan you're solving or to satisfy the audience that you're satisfying.
Alp Uguray (08:40.376)
So.
Bernard (09:05.986)
It's very important to in an environment that allows you to be in a mindset where you can dedicate yourself fully to that problem and not feel the urge to give up or to question your motivations for the problem. And with the weird thing is that everybody is living this entrepreneurial life. And even if they're not, they understand it. So even when you hang out with people that may be doing sales in, I don't know, Oracle, they understand, yeah, that's the startup life. Everybody's like this. Whereas, you know, sometimes I meet my friends who may be working, I don't know, in
Alp Uguray (09:16.802)
Mm-hmm.
Bernard (09:36.194)
know, some sort of financial consulting, which also work, but for them they say, okay, well, it also work a lot. What do you mean? Like you're going to come out tonight. It's like, wow, it's different.
Alp Uguray (09:36.632)
In a way that it allows you to actually feel motivated even though like you stay in and heads down and working and to that point so what we true like how you and Tony like your researchers and then you said you stumbled upon stack AI like how did that journey come true?
Bernard (09:46.764)
while you're Absolutely.
Bernard (10:03.266)
Yeah, that's a very good question. It's a very fun story. So Tony and I essentially like became friends at the beginning of MIT. We actually became friends in an AI conference in Montreal where we both met. It was funny. I always tell people go to the poster sessions in the AI conferences because that's I Tony, the poster session. I saw his poster and I told him, I heard your name. And we just kept talking, became friends. And we saw each other occasionally here in Cambridge. Then after COVID hit,
Alp Uguray (10:07.373)
Thanks, guys.
Alp Uguray (10:30.21)
We're good.
Bernard (10:30.978)
Everybody was locked down. Nobody wanted to go home. A lot of people were working remote. A lot of our friends had graduated and left. So people weren't very interested in hanging out and you know, just talking and meeting up with other people. Except for Tony and I. So we became very good friends over COVID. We became super close. We also realized we were neighbors. So that happened a We're like a block away from each other, which here in Cambridge is pretty small. And we both knew and realized very quickly, we have both this passion for algorithms, mathematics, you know.
Alp Uguray (10:33.398)
We'll talk to our friends about it.
Bernard (10:58.55)
He was very interested in robotics and computer vision, and he was very interested in planning algorithms and optimization. Through that path, we became very good friends. We both knew we both had the same background. We both talked about, eventually after the PhD, we probably both want to begin a startup and we're on different paths. After a while, we just said, hey, you want to begin a startup, I want to begin a startup, let's begin a startup. Let's begin startup. We kind of do it. Probably, we started taking seriously this idea in
Alp Uguray (11:04.332)
it.
Start up. Yes, we'll start up together.
Bernard (11:28.838)
May 2022 when we were about to graduate with PhD, it was becoming more real. It's okay, we have six months before we graduate with our PhDs or we could graduate in six months. Let's actually take it seriously. What are we going to do? Okay, because we've got time playing around with ideas and whatnot. We say, for sure we know that AI is a wave that is coming and we both have invested significantly in this field. We know what are the problems that people building tools with AI face, especially when comes to bringing AI to industry. Now, let's talk about motivation.
has this entire breadth of power to be a system of analysis, to bring insights, to automate tedious and unnecessary work. And the reality is that most AI at the time was just living in a Jupyter notebook. Only a data scientist that could go collect a data set, train a model, and then deploy that model in some complicated cloud could really apply it. And because of that, AI was always limited to that data scientist research world, analytics world of the enterprise. And it was never going.
Alp Uguray (12:11.544)
Exactly.
Bernard (12:25.794)
beyond that scope, very rarely, sometimes very many application-specific setups. We said, okay, maybe we should focus on solving that problem, which is a big thing. And we said, okay, how we can do towards that? Our first question was, okay, why are people struggling to bring more AI models? And we thought about it from perspective of data scientists. We weren't very excited about foundation models at the time, but we knew that there was potential in fine-tuning them, training them, and having the ability to build your own models. We actually had lot of experience before building foundational computer vision models.
Alp Uguray (12:39.16)
Mm-hmm.
Bernard (12:56.022)
Yeah, we were trying in past. So we said, OK, this is for sure a problem. And the biggest problem for us was data. How do you get good data? And at that time, we found Stack AI as a tool for taking a data set of images or text or whatever and cleaning it up and giving you the correct, the best labels to train your model or fine tune your model.
Alp Uguray (13:14.196)
interesting. So it was more like the preparation tool before. they did. Yeah.
Bernard (13:17.056)
It was a preparation version control. It was a very management tool. It was focused towards training models. We took the idea, we built a prototype. I reached out, we reached out to a bunch of people. We got a few clients and from those few clients, we actually got people trying and they liked it. So we said, okay, we're about to get it. It's about the Y Combinator. So we applied to it. It's actually the second time we applied before it was kind of like us playing, but we applied seriously. They said, okay, we actually want to do this idea. We have customers. We are very convinced that this idea of anything is just going to grow. And then we applied, got in.
Alp Uguray (13:31.992)
How was that process? So you apply it to IC and then what were some of the things that they asked in interview?
Bernard (13:45.922)
I just said, okay, we got in. Maybe we should graduate with a PhD.
Bernard (13:56.13)
The initial questions they asked you in the interview are very straightforward. said, why are you building? Why do know this is a problem? Who are your customers? Which customers have today? Is anybody trying this? How much time before the product is ready to actually grow and scale? Very basic questions, which I mean, you can fall very quickly in the first of them. Why are you building? And then second one, who are your customers? How do know they want this? It's very easy to fall through. The waiting time is 10 minutes, so it's very easy. We got through it.
Alp Uguray (14:11.32)
Yeah.
How long did you have like until the fear?
Bernard (14:26.522)
And then we said, okay, we're going to YC. Sure, we're still in the PhD and YC is in two months. We had to have two months, but you know, it wasn't going to be... Me, I was in a position where I had most of my research set up. I got very, I would say lucky, know, thankfully I had a few breakthroughs earlier in my PhD that allowed me to kind of make my final year much more about aggregating the work that I had done into kind of like some key results.
Alp Uguray (14:40.236)
That was it.
Alp Uguray (14:46.776)
Yeah.
Bernard (14:55.317)
rather than kind of like still trying to pursue those breakthroughs. That helped me kind of like accelerate my graduation more. Tony had some work to do, so he had to work a lot to finish. I kind of think this is okay, but I will code the first version on the platform while Tony finishes his PhD and works on getting the branding on the page. Then we'll switch and we'll switch. And you know, we both sprinted to finish the PhD, finished December, two days apart, December 15th, something like December 15th and December 17th. 19th actually for him, 19th. And then we said, wow, we're done. We took 10 days off.
Alp Uguray (15:14.594)
Two days
Bernard (15:25.305)
because it was New Year's and I got, you like, no, no, like, we also had our lease here still, so then they saw. And then we moved to California. We took a flight, arrived to California and started working in Stack AI right away.
Alp Uguray (15:38.988)
And how was that like? You came right there and then you had a product that's more on to prepare a foundation model with better data labeling. So what made it more true to YC program for you guys to realize that, wait, let's maybe pivot it a little bit more or take a different direction.
Bernard (15:44.985)
Yeah.
Bernard (15:48.547)
Yeah.
Bernard (15:56.185)
Yeah. So at that time we were serving more like data teams and startups. we took a big bet at the beginning, we said, okay, there's an extreme demand for fine tuning language models and building custom language models. I ChaiGBT had just come out and people were excited about it and had to make it work. So said, okay, let's gear the product a bit more to focus on language models and NOP and to work towards this task. And we started getting a lot of customers that wanted this, or they thought they wanted this.
Alp Uguray (16:09.528)
Let's.
Bernard (16:23.545)
So we really, we took the product, we shifted towards language models. We made an announcement online, know, hacker news, know, internal ways in there or not. And we got to meet a few companies wanted to try it. We reached out to people, hey, you we have this tool and we got like our first four or five customers on an idea that kind of like that we just made this NLP idea. And they said, okay, this is what we're pursuing. Let's go deeper. Okay. So what do they want? I started working with them at the beginning because the product is kind of being built while they're paying for it. You're always like very on top of them and you're kind of guiding them and you're almost like doing consulting.
Alp Uguray (16:24.184)
Yeah.
Alp Uguray (16:35.349)
Okay.
Bernard (16:51.693)
That's normal because you're giving them a product that can have works and you're optimizing the product for what they want. But through that process and those first two months with Combinator, a bigger decision we had was that most of what this company wanted to do was to take some foundation model and enrich it with the tools and data that they have in their business in order to automate or help access information in some process that they had. And really towards that, you don't really need
Alp Uguray (17:06.84)
Yeah.
Bernard (17:20.643)
to achieve that, you don't really need to fine-tune a model. In fact, fine-tuning a model will ruin a model towards this task. You can steer it towards following a particular language or a particular format in the output, but you cannot really solve this task by fine-tuning a language model. What really was important was to compose logical structural frameworks and workflows where you will retrieve data, hand it to a language model, it became ragged at the time. Back then, people didn't call it ragged because there was only one paper about the topic, and it wasn't very popular.
Alp Uguray (17:49.3)
Yes. Almost the, because I think in the beginning, especially they're like the different type of users, like there's the.
Bernard (17:50.753)
And he was talking anyway, I say it was very funny. And I guess the time came up and
Alp Uguray (18:04.01)
is the business users who are more prone to like a local style app and then they're more technical users who will maybe feel more comfortable in coding. Yeah. Then when you were getting those first customers in, how was the persona like? Yeah.
Bernard (18:13.677)
Yeah.
Bernard (18:20.077)
Yeah, that's actually something we tried a lot. Once we realized that what made sense was having more of a workflow builder tool that could achieve this task and realize its conversion, we got a lot of interest. The moment we released our first launch online, we got 10 customers in a week. And we said, OK, this is definitely worth pursuing. The set of customers that we initially got were more on the executive business side of things. And that was very interesting because we said, OK, there's definitely a potential here for the enterprise to build internal tools.
and a potential there for the startups to build prototypes and a potential for the small businesses to automate small processes. And I very sure which ones to pursue. The first batch of customers were in the startup and enterprise side of things. And we saw the differences there. were developers that were prototyping something. There were companies that building some internal tool. Slowly, we found that the business people that build an internal tool and use it actively, once they have this tool working and they gain value from it, they never leave it. Developers, when they build a prototype and they release it,
Alp Uguray (19:01.272)
That's all.
Bernard (19:17.913)
They moved to code it themselves. So, know, this doesn't really make much sense to build a no code tool for people that want to code. through that experience, okay, our focus should be as a business user. And then the other day was, okay, do we want to go to enterprise? want to go to small business. And through a process of more than
