I am a PhD student from Munich, Germany, working on deep learning for 3D medical data in the form of finite-element meshes and point clouds. I employ graph-convolutional neural networks (GCN) and use symmetry (geometric deep learning) and physics-informed deep learning to boost their performance.
Additionally, I am interested in the mathematical foundations of deep learning through the application of functional analysis and identification of the interplay with dynamical systems and partial differential equations.
I am excited about (super)computers, especially Linux-based systems and clusters. Consequently, I wrote my master's thesis on Hessian-based optimisation of deep neural networks in the context of high performance computing (HPC).
Open master thesis project: I am currently looking for a student to work on the interface of neural ODEs, optical flow and physics-informed neural networks (PINN) for blood flow in arteries. If you like programming (Python, JAX), deep learning and PDEs feel free to reach out to me!