About me

I am a senior-year Doctoral Researcher in Computational Psychiatry in the Institute for Human-Centered AI at Helmholtz Munich and the Ludwig-Maximilians-University Munich under the supervison of Dr. Eric Schulz. I am eager to use my skills to make this world a slightly better place.

Research Background
My PhD work was mostly focussed on developping reliable computational models of human decision-making and studying how these processes relate to symptoms of anxiety and depression.
Now, my work can be largely summarised as psychiatry for AI and AI for psychiatry.
Psychiatry for AI: I evaluate the behaviour of Large Language Models (LLMs) in a range of psychiatric contexts, such as users discussing their psychiatric struggles over longer timespans.
AI for Psychiatry: I investigate how LLMs can be used in psychotherapy research and practice. I build LLM applications for trialling different psychotherapy approaches as well as LLMs that can be used to provide psychotherapy feedback.

Where my passion lies
I want to make sure that AI development and deployment is proceeding in ethical ways. I am therefore working on building evaluation pipelines and continuously learning about AI Safety and my role in this increasingly urgent field.

Outside of work
I really like bouldering, hiking, sailing and generally being outdoors. I have a lot of plants and I do my best to keep them alive. I’m a terrible dancer and even worse singer but I deeply enjoy both.

News

• March 2026: I was accepted to attend the ML4Good Technical AI Safety Bootcamp in May. Can’t wait!
• Febuary 2026: I attended the annual meeting of the International Association for Safe and Ethical AI (IASEAI) in Paris. So many insightful conversations!
• January 2026: I gave my first ever lecture to LMU medical students on challenges and opportunities of using AI chatbots in Medicine.
• July 2025: The paper A foundation model to predict and capture human cognition was accepted for publication in Nature.
• July 2025: I am attending the third edition of the Computational Psychiatry Conference in Tuebingen and presenting a poster on a proof-of-concept study of the usage of open source LLMs in psychological talking therapy
• June 2025: Our paper on the reliability and validity of exploration task parameters was accepted for publication in Scientific Reports. You can find the full article here
• October 2024: A new preprint, Centaur: a foundation model of human cognition, is out.