Fellowship

The Fellowship Program connects technical students at Lund University with nonprofits that have real problems software can solve.

Each semester, ten selected students spend fifteen hours a week on a single project for an organization. A team might fine-tune a language model on domain-specific data, build a predictive model on years of untouched records, or replace a pipeline that runs on spreadsheets and manual effort. Jonathan Cederlund, who previously did research at Caltech, supervises the work. Anthropic sponsors with cloud credits.

Everything is open source from the first commit. In January and June, teams present their work to partners and sponsors. At the end of each semester the organization receives the code, the documentation, and a proper handover. Projects that need more time run another semester.

For the organization, the work is pro bono. In return we ask for a contact person who can give two hours a week of guidance and data access, and that you cover direct costs such as API fees or data licensing.

We’re looking for scoped AI and ML projects now. The next ones start the first week of September.