I joined to execute. I ended up deciding.
18 months, one expensive migration, and what nobody tells you about growing inside a startup.
You're in a meeting. The question isn't which technology to use — it's how to build a platform that works without breaking the budget. Which AWS services are worth paying for. Which ones to replace with your own solutions to bring costs down. Until that moment, a dashboard had been making all of those decisions for you.
You took the role anyway. You asked to be in this room. And you spoke.
That's where it started.
Until July 2024 I was a backend dev who deployed on Render or Railway. Push the repo, set the env vars, get a URL back. That was "infrastructure" to me — a managed service that hid everything I didn't want to think about. Then I joined Brault, and the first month they handed me the central platform architecture. AWS, microservices, real decisions, zero roadmap. No platform to abstract the complexity away.
How do you close that gap? I went into the AWS console and started touching things. Service by service, as I discovered I needed them. I read the fine print, read the documentation, tested, broke things, and read more documentation. No course, no certification, no one else's infrastructure to copy from — just direct investigation, understanding each service well enough to know whether to use it, build around it, or replace it with something cheaper.
When I joined Brault, the stack was already chosen. AWS, Node, MongoDB. I was there to execute, not to decide. Two weeks in, looking at the domain models, I knew something didn't fit — the entities had obvious relationships that MongoDB was going to force us to wire by hand. I raised it. It wasn't made a priority. The team had other things to ship, and that's how startups work: under real constraints, not ideal ones. So I kept patching the workarounds. Months later we migrated everything to a relational database. Time, risk, full service rewrites. Costlier than anyone had estimated.
But that migration was the first time I wasn't just implementing a plan — I was owning an outcome. There's a difference between raising your hand and carrying the solution to the end.
Growing inside a startup doesn't happen in one moment. It's a sequence: a project designed from scratch and improved over time, a technical crisis resolved before anyone had to ask, an architecture decision that nobody questioned. And then, at some point, the partners started coming to me instead of the other way around. That shift — from waiting for approval to being the one people ask — is the actual transition. It doesn't have a name or a date. You notice it only after it's already happened.
At some point I stopped working for Brault and started working in Brault. Shared ambition. Real trust. Decisions that matter beyond the next sprint. It's still constant — there are still migrations to do, crises to solve, decisions to argue for. The path isn't cleaner now. I just know the terrain better.
I didn't wake up one morning as an architect. I became the person people called when something mattered.
That's all there is. No certification for it. Just decisions, time, and the project you choose to believe in.
Has something like this happened to you? Write me at hello@sebastianfermanelli.com.