Radovan Matula is a PhD student in the field of theoretical condensed matter physics.

His research interests include ab intio modelling of materials, primarily with density functional theory. For his PhD research he is exploring the use of machine learning techniques to improve sytem-size and temporal scaling of DFT.

He started his work with DFT during his bachelor thesis, when he worked on modelling the band structure of CsPbBr3 perovskite in tandem with experimental research. He continued to his master’s combining DFT with machine learning to study the halide exchange in lead mixed-halide perovskites. He continues to build on his previous experiences during his PhD.

Radovan is a PhD student in the Modelling and Theory of Materials Group since 2025.

DOK-NPOO-2023-10-2165

Within the HrZZ DOK project his research focuses on machine learning for atomistic simulations, with the main goal of his PhD being the development of universal machine learning interatomic potentials (MLIPs) for molecular crystals. He works on developing a workflow which combines active learning strategies with the MACE architecture to efficiently generate high-quality training data and ensure the transferability of the resulting models across diverse crystal systems.