Workplace/Department: Saarland University, Chair of Data-Driven Simulation and Analysis in Material Science, Subject Area: Foundation Models and Agentic AI for Physical Systems

What we work on:

Our group develops LLMs and agentic AI systems for scientific discovery, engineering and physical systems. We investigate how AI can reason about scientific problems, interact with simulation software and support the design of complex physical systems. The group is jointly embedded at Saarland University and DFKI in Saarbrücken.

Current Research Directions include:

• AI for scientific computing – neural operators and learning-based surrogates for physical systems
• LLMs and scientific agents – large language models that autonomously reason, plan and execute scientific workflows
• AI for engineering design – LLM-driven agents for autonomous optimization and design of materials, structures and processes
• Multimodal scientific AI – integrating language, images, sensor data and simulations
• Alignment for scientific AI – scalable oversight, verification and control of LLM-based agents
• AI for materials discovery – generative models, active learning and autonomous discovery of novel materials

Your potential responsibilities:

• Pursue an independent research project, develop your own research line, within one of the research directions above or a closely related area
• Develop, train and evaluate modern ML models on GPU/HPC infrastructure – including graph neural networks, transformers, neural operators, diffusion models and foundation-model-based systems – and, where relevant, combine them with numerical simulation and scientific computing workflows
• Publish at leading venues (e.g. NeurIPS, ICML, ICLR, AAAI, npj Computational Materials) and present your work at international conferences
• Contribute to collaborative research projects with our close partners at EPFL and at Hamburg University of Technology and help shape future research initiatives and funding proposals
• Support teaching, co-supervise students (PhD/Master/Bachelor, as appropriate to your level) and collaborate actively across Saarland University, DFKI, Saarland Informatics Campus, and international partners

Your academic qualifications:

• Doctoral degree / PhD in one of the above disciplines, with a demonstrated research record at the interface of AI and science
• Language skills (according to GER): English C1