About

I’m a staff-level engineer based in Berlin, originally from Mexico City. Over the last 10+ years I’ve moved from building platform systems at scale, to consulting on other companies’ architecture, to founding my own products. I was a Senior Engineer (L6) at Unity, where I worked on platform systems serving roughly a billion requests per month, then a Lead Engineer at BCG Platinion, helping clients untangle infrastructure and organizational problems. These days I’m building two things: Escribano, a local-first memory system for AI agents, and raus.cloud, an infrastructure consultancy.

How I think

Decisions belong in writing. I use ADRs for anything that will matter in six months. A decision that only exists in a meeting or a chat thread doesn’t exist. Writing forces the trade-offs into the open and gives future engineers — including me — the context to change course without archaeology.

Threat modeling comes before scaling. Before I make a system bigger, I think about how it fails and who might misuse it. Scaling a system with unclear trust boundaries just scales the incident. Cheap, early threat models beat expensive, late security reviews every time.

SLOs are contracts between teams. An SLO is not a monitoring dashboard; it’s an agreement about what users of a service can rely on, and what the owning team can stop worrying about. Error budgets make reliability and velocity trade-offs explicit instead of emotional.

Fire yourself. My job on any team is to make myself unnecessary. That means documenting what I know, automating what I do, and mentoring people until they don’t need me. A team that depends on one person is a team I haven’t finished building.

Write architecture for machines too. I now write architecture documents and ADRs so that AI agents can read and act on them — explicit invariants, clear boundaries, no implicit knowledge. Agents are becoming contributors; they deserve the same clarity we owe our teammates.

What I’m looking for

I’m interested in staff engineering roles where writing, systems thinking, and mentoring matter as much as shipping code. I also take on a small amount of advisory work through raus.cloud, mainly around infrastructure, reliability, and AI agent architecture.