Postdoctoral researcher at the ELLIS Institute Tübingen (COMPASS group), working on trustworthy, safe, and effective reasoning in language models and agentic systems.
