Erik J. Bekkers is an Associate Professor in Geometric Deep Learning at the University of Amsterdam, affiliated with the Machine Learning Lab (AMLab). Previously, he served as a post-doctoral researcher in applied differential geometry at the Technical University Eindhoven (TU/e) Department of Applied Mathematics, following completion of his PhD cum laude in Biomedical Engineering at TU/e. His educational background includes: PhD in Biomedical Engineering, Technical University Eindhoven (cum laude) Dr. Bekkers' research centers on geometric deep learning principles where data representation preserves physical-world geometry and symmetries. He develops group-equivariant architectures and explores structure-preserving learning through equivariant operators, dynamical systems on manifolds, and geometric algebra. His work bridges theoretical frameworks with applications in medical imaging, computational physics, robotics, and generative modeling of non-Euclidean data. His scientific recognitions include: MICCAI Young Scientist Award 2018 Philips Impact Award (MIDL 2018) NWO VENI grant: "Context-Aware Artificial Intelligence in Medical Image Analysis" NWO VIDI grant: "Neural Ideograms: Shaping AI with Geometry-Grounded Learning" As principal investigator of NWO grants, he leads research on geometry-grounded representation learning while co-organizing the ICML'24 GRaM workshop to advance community collaboration. His GitHub repositories demonstrate active software development in equivariant neural networks. At the AMLab, his team investigates geometric latent variable models, physics-informed neural networks, and methods for generating geometric objects on manifolds, maintaining strong ties to both theoretical foundations and real-world applications.









