Projects
Focus
A few things I keep coming back to.
AI security
I find prompt injection and system-prompt extraction issues in LLM apps and help fix them.
Reverse engineering
I take apart AI tools to see what they send to the model: system prompts, agent loops, routing.
Open source
Most of what I build is open source, and I keep maintaining it after release.
Web
React, Next.js, and TypeScript. This site and most of my projects are built with them.
Elsewhere
Email me at lucasvalbuena@pm.me or find me on GitHub, X and LinkedIn.