Michael Obersteiner

Dr. Michael Obersteiner is a researcher at the Technical University of Munich (TUM), affiliated with the TUM School of Computer Science and Information Technology and the Department of Computer Science. He is part of the Chair of Scientific Computing in Computer Science (SCCS), led by Prof. Hans-Joachim Bungartz. His research focuses on high-dimensional numerical methods, sparse grids, the combination technique, molecular dynamics, and machine learning applications. He has contributed to the development of scalable algorithms for exascale computing, including the DisCoTec and sparseSpACE frameworks.

Education: Michael Obersteiner began his studies in Biochemistry in 2010 before switching to Computer Science in 2011. He completed his M.Sc. in Computer Science in 2016 and is currently pursuing his doctoral dissertation with a focus on high-dimensional numerics and the sparse grid combination technique.

Research Interests: His work spans molecular dynamics simulations, high-performance computing (HPC), spatially adaptive sparse grid methods, and applications in uncertainty quantification, machine learning, and plasma physics. He has developed open-source software tools like DisCoTec (for exascale PDE simulations) and sparseSpACE (for adaptive sparse grid operations).

Teaching: Obersteiner has been involved in teaching numerical programming, scientific computing, and discrete structures at TUM. His roles include leading exercises for courses like 'Algorithms for Scientific Computing' and 'Numerisches Programmieren.'

Projects: He leads projects such as DisCoTec and sparseSpACE, which focus on scalable sparse grid techniques for high-dimensional problems. His research also includes applications in plasma turbulence modeling, hydrology simulations, and machine learning with density estimation.

Grants and Collaborations: His work has been supported by projects like EXAHD (funded by DFG/SPPEXA). He collaborates with institutions like BGCE, CSE, and IGSSE. Open student projects are available in areas like adaptive quadrature, machine learning with sparse grids, and uncertainty quantification.