Welcome to the homepage of Thomas Hamelryck, a researcher at the University of Copenhagen with a full professor position in High dimensional biological data analysis / Machine learning shared at the Department of Biology (Section for Computational and RNA Biology, SCARB) and the Department of Computer Science (Programming Languages and Theory of Computation section, PLTC).

I am a specialist in Bayesian modelling and probabilistic machine learning. Currently, my research mostly focuses on probabilistic machine learning applied to protein structure prediction and protein evolution. I am particularly interested in the use of deep probabilistic programming, making use of the deep probabilistic programming languages Pyro and Numpyro, and the application of directional statistics to represent non-Euclidean data. My group also contributes to the development of Pyro and Numpyro.

My group's research featured in journals such as PNAS, PLoS Computational Biology, Bioinformatics and Molecular Biology and Evolution, as well as in leading AI conferences such as ICML and ICLR. In addition, I am the author of the Bio.PDB module of Biopython, a standard open-source library for computational structural biology, used widely in both academia and industry. This research has applications in biotech and medicine: a deep probabilistic model of local structure was for example used in multivalent vaccine design by a Danish biotech company.

(Picture credit)