About me

Hey, I'm Antonio 👋

I work as an AI engineer, building agent systems and trying to figure out how to make them actually reliable. Right now I'm deep into DSPy (way better control than LangChain's magic), Python tooling (uv, ruff, ty), and the whole problem of agent verification; how do you make sure they do what you want without just hoping for the best?

I write here about what I'm learning, what works, what doesn't, and opinions on the tech world. No fluff, just real stuff.

What I care about

AI agents that actually work. Not demos. Not tutorials. Real systems you can trust. That means verification, enforcement over advice, and understanding the stack deeply enough to build your own solutions.

F1 and simracing. Big Alonso fan. I race in VR on iRacing with a full Fanatec setup.

Thematic investing. Infrastructure, utilities, electrification. I like understanding the thesis behind what I buy.

Cooking with intent. If I'm making something, it's worth doing right. Lasagna? 3-hour braise. I organize recipes in Cooklang with reusable bases.

How I work

I'm pragmatic, hands-on, learn by doing guy. I like building systems, not following recipes. I want to understand how things work under the hood so I can make them do what I need, not just what the docs say.

Background

Before AI, I was a professional beatboxer (166K on YouTube), competed internationally, went on Spanish TV. That taught me how to perform and improvise under pressure.

I'm also an aerospace engineer from UPM (Madrid), spent two years building Formula Student cars.