Everything I've built, grouped by purpose.

Joining as engineer. Career advising for the most pressing problems — now from London.

A serious publishing venue for alignment research. Cofounder; currently sole engineer.

AI safety's career platform. One profile, smart recommendations, automatic recruiter matching.

The Buenos Aires AI Safety Hub. Cofounded; handing off as I move on — possibly staying on as advisor.

Red-teamed ASI governance frameworks — 1st place at Apart Research hackathon, parlayed into a month-long governance fellowship.

A practical benchmark robustification toolkit for CBRN safety evals — 2nd place at Apart Research hackathon.

Multi-agent framework that compresses AI governance wargames from years to minutes. Apart Research hackathon submission.

Demo day site for the Kairos cohort at SPAR. Commissioned, built in December.

Personal site for an EA therapist. Clean one-pager, quick turnaround.

Research lab site for FAIR at the University of Buenos Aires — papers, team, projects, built on SvelteKit with a git-backed CMS.

Token-efficient CLI for Google Docs. Built for AI agents — read, write, and collaborate with change detection and conflict prevention.

macOS speech-to-text that streams local-model transcription straight into a markdown file, in real time.

Autonomous milestone executor — runs discuss → plan → execute workflows end-to-end with parallel Opus + Codex validation at each boundary.

Reusable Python wrapper around the Claude Agent SDK — retry, timeouts, and workarounds, so you stop writing the same glue twice.

Prometheus had the right idea. Paused post-80k — still +EV, resumable later.

AI safety documentaries. 20M+ views. Research, ops, and production — left in February 2026.

Photos that prove they're real. C2PA cryptographic proof + depth data.