Ignacio Roldan Fernandez PRO

Aerospace engineer transitioning to technical AI safety research. Currently working on LLM honesty, deception, evaluations and context-dependent linear representations. Interests include AI control, scalable oversight and mechanistic interpretability.