My research mainly focuses on gaining better understanding of open challenges in generative modeling, representation learning and robust decision-making. My background is in physics with a specialization in neuroscience.

I have been working on representation learning with generative models, on invertible neural networks and normalizing flows, on robustness under distribution shift as well as on the connection between invariance and generalization. Currently, I mostly think about ood generalization, (private) representation learning, algorithmic bias and the consequences implied by methods and common practices in AI/ML research. Also, I run my own diy record label (check other projects tab).