AI for Scientific Discovery in Physics

Our group builds artificial intelligence systems for discovering new concepts, experiments and ideas in physics.

To accelerate this effort, we need your help!

We have several fully-funded open Post-Doc positions at the University of Tübingen.

A list of concrete potential projects:
• Development of modern auto-differentiation (JAX-based) physics simulators for the discovery of new physics experiments (example here)

• Developing and testing state-of-the-art AI-driven exploration, optimization, and search algorithms in extremely complex and enormously large spaces motivated by physics and chemistry

• Agentic frameworks (e.g. LLMs with tool-use) for closed-loop idea generation for physics (example here)

Other projects are certainly possible too. In general, we believe that building autonomous scientific systems is not just a technical question, but requires understanding and insights from the philosophy of science – see e.g. here.

If you are excited to use artificial intelligence techniques for scientific discoveries in physics, send us your application, including a CV, a short motivation letter, the names & contact of two potential references to mario.krenn@uni-tuebingen.de. The opening will remain valid until 28.10.2025 (unless filled before).