Hannes Stärk
PHD Student at Massachusetts Institute of Technology
Cambridge, Massachusetts, United States
Summary
🤩
Rockstar🎓
Top SchoolHannes Stärk is a PhD student at MIT CSAIL specializing in graph and geometric machine learning with a focus on self-supervised methods for proteins and small molecules. He holds an M.Sc. in Informatics from TUM and brings eight years of experience across research internships and applied ML roles, including work on molecular docking (contributing robustness and preprocessing fixes to the widely used DiffDock implementation). At MIT he is advised by Tommi Jaakkola and Regina Barzilay, blending rigorous theory with practical pipeline engineering for biomolecular problems. His background spans teaching, web and cloud engineering, and operations research, which helps him bridge modeling, data engineering, and reproducible research. Colleagues describe him as approachable and eager to collaborate on both deep technical challenges and broader scientific questions.
8 years of coding experience
1 year of employment as a software developer
Bachelor of Science - BS, Informatics, 1,68, Bachelor of Science - BS, Informatics, 1,68 at Bundeswehr University MunichDoctor of Philosophy - PhD, Computer Science, Doctor of Philosophy - PhD, Computer Science at Massachusetts Institute of TechnologyMaster of Science - MS, Informatics, 1.2, Master of Science - MS, Informatics, 1.2 at Technical University of Munich
Github Skills (14)
docking10
machine-learning10
computational-biology10
python10
debug9
data-preprocessing9
error-handling9
debugging9
rdkit9
file-handling8
file-processing8
fileio8
evaluation8
file-access8
Programming languages (3)
Github contributions (5)
gcorso/DiffDock
Oct 2022 - Jan 2023
Implementation of DiffDock: Diffusion Steps, Twists, and Turns for Molecular Docking
Role in this project:
ML Engineer
ML EngineerContributions:29 commits, 2 PRs, 27 pushes in 3 months
Contributions summary:Hannes primarily contributed to improving the data preprocessing and evaluation pipelines for the DiffDock project, a molecular docking implementation leveraging diffusion models. Their work focused on fixing bugs in the ligand embedding preparation and inference processes. They also added logging and improved error handling within the dataset loading and evaluation scripts, thereby enhancing the robustness and usability of the model.
dockingnon-euclidean-geometryscore-based-modelsmolecular-dockingmachine-learning
HannesStark/EquiBind
Feb 2022 - Oct 2022
EquiBind: geometric deep learning for fast predictions of the 3D structure in which a small molecule binds to a protein
Contributions:30 commits, 5 PRs, 31 pushes in 7 months
moleculebindsgeometric-deep-learningdockinggraph-neural-networks
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.