Dimitris Loukrezis

Short Bio

I joined the Scientific Computing group at CWI Amsterdam in March 2025 as tenure-track researcher. Previously, I worked for Siemens AG in Munich as research scientist for modeling, simulation, and optimization of digital twins since 2021. I was also post-doctoral researcher at the Institute for Accelerator Science and Electromagnetic Fields of TU Darmstadt, where I started and led the Scientific Machine Learning research group (full time in 2019-2021, part time in 2021-2025). I obtained my Dr.-Ing. degree from TU Darmstadt in 2019, working on high-dimensional uncertainty quantification for electromagnetic fields. I also hold an M.Sc. in Simulation Sciences (RWTH Aachen, 2014) and a Diploma (integrated masters, B.Sc.+M.Sc.) in Electrical and Computer Engineering (NTU Athens, 2012).

Research

My research touches upon and tries to connect three main areas of interest, namely, uncertainty quantification, surrogate/reduced-order modeling, and scientific machine learning. I am particularly interested in problems and applications suffering from the so-called "curse of dimensionality", as well as in data-driven yet physics-conforming predictive models.

Indicative topics of interest:

• High-dimensional model learning and surrogate modeling.
• High-dimensional uncertainty quantification and sensitivity analysis.
• Structure/physics-preserving dimension reduction.
• Physics-conforming (scientific) machine learning.
• Uncertainty quantification for scientific machine learning.

Application-wise, my current focus is on electric machine design in the context of the CREATOR Collaborative Research Center, as well as on the development of digital twins.

For a complete list of my publications, check my Google Scholar webpage.