Martin Michajlow

Martin received his master's degree mathematics from TU Berlin in 2024. His master's thesis on "Equivariant Transformer Networks for Molecular Multipole Prediction" was written in the machine learning group of Prof. Dr. Klaus-Robert Müller under supervision of Thorben Frank, where he subsequently started his PhD journey. His main research interests lie in the application of machine learning to quantum chemistry, more specifically at the intersection of graph neural networks and self-consistent field methods. There he tries to build a principled understanding of the connection between the implicit modeling of the electron density in atomistic graph machine learning frameworks and the all-electron descriptions in density functional theory and wavefunction methods like Hartree-Fock. This entails the functional expressions used for describing molecular and electronic structure in both fields, as well as the mathematical methods for finding optimal solutions.