Bio
AI safety research: LLM instrumentally convergent behaviors (Robert McCarthy, UCL/MATS) and persuasion evaluation (Kellin Pelrine, FAR.AI).
SDE Intern at Amazon (agentic RAG tool).
LLM evaluation at ECOLE Lab; organizer for the Berkeley AI Safety Initiative (BASIS).
UC Berkeley undergrad CS + Applied Math.