I work on the plumbing side of AI safety.

Evaluations, hiring pipelines, publishing venues, training programs — the stuff that’s mechanical until one day it isn’t. Not the research, just the things around it that have to work for the research to happen.

I got pulled in by an ACX grant in 2022. Spent 2023–2024 doing ops at Nonlinear, running their grantmaking network — 50+ funders, 500+ applications, $800K+ channeled — and helped stand up the emergency fund that got another $800K to FTX-affected grantees in three days.

Alongside all of that, spent 2023 through early 2026 at AI Species, an AI safety documentary project that cleared 20M+ views — doing background research, fact-checking, and ops.

In 2024 I cofounded BAISH, the Buenos Aires AI Safety Hub. Since then I’ve also cofounded ASTN, a career platform for the field; started shipping as the sole engineer at the Alignment Journal; and facilitated BlueDot’s AGI Strategy and Technical AI Safety courses.

In 2026 I’m joining 80,000 Hours as an engineer. Moving to London for it.

Most of what I’ve shipped has been an attempt to close AI safety’s biggest talent gap — the engineers, the ops people, the fieldbuilders who make a field run. Or, occasionally, just a thing I kept wishing existed.

I’m Luca De Leo. Buenos Aires → London.

If you want the formal version, here’s my CV.