Jonas Köhler

About

I studied pure math (algebraic topology), theoretical physics and media art (sound installations) for some years, before I eventually entered computer science at age 25.

Here I did my B.Sc. in Media Informatics at Bauhaus-University Weimar (2016) and my M.Sc. in Artificial Intelligence at University of Amsterdam (2018).

During my undergrad I worked on information retrieval and natural language processing (argumentation mining, locality-sensitive hashing) in the webis group (opens in new tab).

During my master I worked with Taco Cohen (opens in new tab), Mario Geiger (opens in new tab) and Max Welling (opens in new tab) on Spherical CNNs (opens in new tab).

I wrote my master thesis during a research internship at Max-Planck Institute for Intelligent Systems, Tübingen (opens in new tab), where I worked with Mijung Park (opens in new tab) on differentially private generative models.

In 2018 I joined the research group of Frank Noé (opens in new tab) at Freie Universität Berlin, doing a Ph.D. on generative models for sampling molecules in thermodynamic equilibrium. Here I mainly worked on Boltzmann Generators (opens in new tab) and extensions (see my scholar for more work on this).

In January 2023, I joined Microsoft Research AI4Science, Berlin, as a Senior Researcher. Here I mainly work on Deep Quantum Monte Carlo techniques and sampling.