Assistant Professor for Scalable and Efficient Machine Learning, TUM Campus Heilbronn

Markus Götz is an Assistant Professor for Scalable and Efficient Machine Learning at the Technical University of Munich (TUM), Campus Heilbronn. He also continues to lead a research and consulting group as Group Leader for Helmholtz AI Energy at the Karlsruhe Institute of Technology (KIT). He received his PhD in Computational Engineering from the University of Iceland, carried out in close collaboration with the Jülich Supercomputing Centre (JSC), with a thesis on scalable data analytics in high-performance computing. He specializes in distributed and parallel machine learning on supercomputers, covering topics from scalable clustering and classification to large-scale neural network training. He also has a long track record of research software development: he is lead developer of the HeAT (Helmholtz Analytics Toolkit) distributed tensor framework, and a core contributor to Propulate, an asynchronous evolutionary optimizer, and perun, a tool for measuring the energy consumption of parallel applications. Markus Götz serves as Helmholtz AI Coordinator Karlsruhe and AI Spokesperson for the Helmholtz Open Science Office, and is a Principal Investigator on numerous national and EU-funded projects.

Research

My research focus is on scalable machine learning and data analytics for high-performance computing environments. In particular, I am interested in distributed and parallel training of neural networks, resource-efficient and low-rank model architectures, energy-aware AI, and scalable clustering and classification methods for scientific data. My work spans applications in renewable energy systems, weather modeling, and biomolecular simulation. I care deeply about sustainable and reproducible research software, and much of my work is complemented by open-source tool development aimed at making distributed AI methods accessible and energy-efficient for the wider scientific community.