Thomas does fun and serious things with deep learning models.
Thomas founded MathInf GmbH headquartered in Münster, Germany. From there he helps clients around the globe to achieve more with their AI. His work the entire stack from novel mathematical methods for training and monitoring AIs to optimization of low-level code. His customers range from research institutions to startups to large multinational companies.
He is an author of Deep Learning with PyTorch with Manning Publications. Thomas created TorchDrift, a library for checking whether your deployed models are still running within spec. At more than 150 features and bug fixes to the early PyTorch codebase, Thomas is a one of the most prolific independent contributors.
A mathematician by training, he has a Ph.D. in mathematical modelling from the University of Bonn.