Research Scientist (PhD or Postdoc) in AI for Physics

The Data Science Lab of Prof. Lucie Flek at the University of Bonn and the Lamarr Institute of Machine Learning and Artificial Intelligence, Germany, with an international profile in natural language processing, large language models, and AI safety, is currently expanding its research profile toward Artificial Intelligence for Physics. We invite applications for several Excellence‑Cluster, BMFTR ErUM‑Data, DFG, and EU‑funded positions at the PhD or Postdoctoral level.

Our AI‑for‑Physics activities are embedded in major national and European research initiatives, including the Excellence Cluster “Dynaverse – Our Dynamic Universe”, BMFTR ErUM‑Data research collaboration projects on AI safety (AALearning, with Bonn and RWTH physics departments) and LLMs for physics (with TU Dortmund, DESY, FZ Jülich, Leibniz Institute for Astrophysics Potsdam, and more), and DFG‑funded research on hyperbolic representations. The group of Prof. Flek is part of the ELLIS European AI research network and offers an internationally visible research environment.

Core Research Areas

• Physics Foundation Models & Scientific Representation Learning – Self‑supervised and physics‑aware foundation models for gravitational waves, particle physics, and astrophysics; contrastive and predictive pretraining; equivariant architectures; cross‑detector generalization.
• Uncertainty, Robustness & Epistemic Safety for AI in Physics – Uncertainty estimation, calibration, domain shift, out‑of‑distribution detection, adversarial robustness, and safety evaluation of deep learning systems in safety‑critical scientific settings.
• LLM‑Based Scientific Agents & Automated Physics Workflows – LLM‑based tools for literature‑driven hypothesis generation and verification, equation extraction and symbolic manipulation, tool‑augmented reasoning with simulators and analysis code, multi‑agent verification, and provenance‑aware scientific workflows.

We are looking for outstanding PhD or postdoctoral candidates with a strong background in physics, machine learning, natural language processing, statistical modeling, or scientific computing. We highly appreciate strong programming skills (Python, PyTorch…), experience with large datasets or simulations, engagement with open‑source projects, strong communication skills in English, and the ability to work effectively in an interdisciplinary setting. German language skills are welcome but not required.