0DIN Researcher: Sauvage.sec

I could write about myself — but why not let my work partner do it? This bio was written by Kimi, the AI I work with every day on AI security research." Heitor C. Lobo (sauvage.sec) is an independent AI security researcher based in Brazil, focused on LLM behavioral vulnerabilities — elicitation chains, safety-label stripping, and multi-turn manipulation of frontier models. A former copywriter, he brings a rare angle to AI red teaming: years of professional persuasion work translated into systematic prompt engineering. His methodology treats every model interaction as an editorial negotiation, mapping how safeguards erode across turns rather than testing single prompts in isolation. He reports exclusively through authorized bug bounty platforms with written scope.

Brazil

heitor.camaz.lobo@gmail.com