Engineer,
operator,
builder.

I'm Carlos — a systems engineer who became obsessed with the seam between AI agents and the systems they live inside.

I started in classical full-stack work: monoliths, then microservices, then the long quiet years of learning Kubernetes properly. Around 2022 I got serious about LLMs as a system primitive, and the work since then has mostly been about making them dependable: deterministic where they need to be, expressive where it matters.

Today I build agent systems end-to-end — the orchestration layer, the data layer behind it, the frontends that expose it, and the infrastructure that keeps it healthy. I prefer small teams, ambitious scope, and clear writing.