We are a research group established in September 2026, supported by Novo Nordisk Foundation, led by Svetlana Kutuzova from the University of Copenhagen, The Department of Computer Science.
We are a CS lab developing deep learning methods in domains of generative and probabilistic modeling, uncertainty quantitation and multimodal learning.
Our application areas and collaborators come from various areas of life science. Currently, we are most interested in empowering high-throughput analytical chemistry techniques, like finding chemicals in vast amounts of wastewater. But we also like microbes and all things multiomics. Don't hesitate to write to svegal@di.ku.dk to learn more.
Our code practices
We aim to keep things simple but organized.
- When starting a new research project, create a repository within the organization.
mainbranch is protected, contributions happen through pull requests. See an example- Give your branches descriptive names with prefixes, like
feature/peak_segmentationorfix/lockspray. It is usually good if they are short-lived and focused on a specific thing. - The exceptions to keeping it small are larger endevoures, like refactorings. Use the
refactoring/prefix and make sure your colleagues know it's happening. - Sometimes when submitting to a journal the version of the code at submission will have to stay frozen in its own branch, like
submission/nature_07082026. In an active project, it would mean that the code will inavoidably diverge frommain. This is just how it is. Do not keep adding things to the submission branch after submission. - It is a (very) good idea that your PR is reviewed by a colleague, but when working on many research projects, we are sometimes uncoordinated and in rush, so the reviewing itself is not enforced.
- Add basic pre-commit hooks so PRs arrive clean. Get inspired by, e.g. this repo
Common sense comes first, so if you think your project should run differently, talk to your colleagues.
