I joined Tako something like a year ago, right after I came back from SF for the first time and was looking for a fast-growing company. What I saw at Tako were founders who had built stuff before and an engineering team with tons of experience. Since my goal was to learn the things I'd need to become a founder, it was a no-brainer: this was the perfect place.
The bad part of joining Tako was having to drop out of university and move from my hometown to São Paulo (Brazil's biggest city by population). At the time, I honestly didn't care. I was so motivated about joining a seed-stage company backed by serious funds, and about learning from the team, that by the time I signed the offer I'd already bought my one-way ticket to São Paulo.
The first months were insane. I learned what serverless meant, learned a lot about AWS, went to events to understand more about the domain (payroll and Brazilian labor laws), and for the first time I felt like I was working with a world-class team. I was also having frequent 1:1s with Fernando (co-founder and CEO) that made me feel so much in the game: sharing my points of view (often contrarian) on the direction the company was heading, brainstorming what to build next, figuring out who we should hire. That made me feel once again that I was meant to be a founder. Fernando became a mentor, an example of what a great leader is, and not because he was always right, but because he was humble enough to know when he wasn't and needed help.
A few months went by and I got my first big project: working with an engineer named Lucas to build our own AI framework. The motivation was that the famous frameworks like CrewAI, LangChain, LangGraph, the OpenAI Agents SDK and many others were full of bugs, lacked personalization, and gave little visibility into agent loop executions and tool calls, and the best framework we knew of (Agno) was only available in Python (our codebase was entirely in TypeScript). So we started a one-month project to build one in-house. That was the first time at Tako that I worked 10-12 hours a day, motivated to ship it as fast as possible so we could actually use it for the product-facing agents we wanted to build. It was a great experience. I accidentally took down our OpenSearch twice, and I found out I could be more productive by running 2 agents in parallel (later I found out the limit for my brain was 6 terminals).
By December, we had figured out that our agents inside the payroll platform weren't the biggest deal in the world, and we noticed there were other opportunities around payroll that we could explore with AI.
In January, we decided the area we'd work on was recruitment. At that point the idea wasn't clear, the ball wasn't rounded enough to kick, and for a few weeks my job stopped being engineering and became product. I hopped on calls with AI recruitment products and talent acquisition people at the biggest companies in Brazil, so I could understand the biggest pains they had in their day-to-day work in that area.
For the big companies it was clear: people didn't have enough hands or time to talk to everyone who applied, and had to use AI filters based on CVs to decide who would move to the next step of the hiring process. So the solution was to build a voice agent that spoke with people for 30 to 40 minutes to do the recruiter's job in the first step of hiring.
After figuring that out, I went back to engineering to build our voice agent, again with Lucas, but also with Pedro (an early-career engineer that I met at Berkeley who had just joined the company), Flávia (a PM who had just joined Tako from Nubank), and Eric (a designer who came from PicPay and was one of the first hires at Tako). It was fun. We worked day and night for 2 months to build the whole product: an agent speaking in Portuguese, connected to the company's ATS, that ran tailored interviews based on each job description and candidate CV. I was happy, working onsite on weekends, jumping on sales calls, and saying the same things to a voice agent 12 hours a day.
In late March the first company started using our product. They were the biggest foodtech in LatAm and had more than 500 job openings a year. After that, 40 other companies started using it. It felt like one of the first times something I'd imagined and built from scratch had reached PMF.
Tako felt like home. Everyone knew me: the way I ran through the hallways, the pizza I ordered every time I stayed late at the office, the music I was most into at the time. Every mentorship session that year with Fernando felt like a giant lesson I had to take notes on. Fernando was a friend, someone who truly believed I could be a great founder someday.
I've always known I wanted to build globally. Even though Brazil is a giant opportunity in specific markets and has incredible engineers, I feel like being in SF at such a specific moment in history, when everything is being completely reshaped by AI, is the best thing I can do with my life right now. It's a bit strange to write that, because not long ago I was the one touring US campuses trying to convince Brazilians to come back home. But I don't think the two are in conflict: I want to be where the frontier is moving fastest.
When I said I wanted to leave Tako to move to SF, Fernando offered me everything to start my own company: money, mentorship, contacts, and a few more months at Tako learning about hiring, product, and sales. For that I will always be thankful to him.
The reason I didn't take it is that I still feel I have a few gaps to close before starting something of my own, and even though starting a generational company is really hard, the challenge of filling the gaps to build it excites me deeply.
So, in the end, I've decided not to start my own company now, but to join an early-stage startup called Sable with two of my best friends, people I'm entirely sure I want to build with. My goal is to get to SF, grow a product a lot, meet incredible people (the ones I've watched on podcasts), study entrepreneurship, tech, design, and how everything is changing, and then build my own company.
Even though this looks like a lot to do before starting my own thing, it won't be a lot of time. The world is changing, and I know I'm not cut out to be an employee forever.
My ultimate goal is to start thinking about an idea while I work, something I'm truly passionate about and important enough that I'd accept eating glass for years if at the end of it I can transform people's lives.
I'd just like to thank Fernando Gadotti, Juliana Jordão, Lucas Monteiro, Sebastián Mejía, Eric Cerqueira, Flávia Cruz, Bruno Tavares, Bruno Cerejo, Vicente Bastista, and a bunch of other people I had the privilege to work with at Tako, who made me believe everything was possible and became so important in my life. I truly hope this isn't a goodbye, but a see you soon.
