Profile

Digital humanities projects rarely fail on technical grounds alone. They fail because the people involved speak different languages: researchers know which questions the sources should answer, developers know what a data model can carry — and what both of them meant gets lost in between.

That gap is where I work. Years of practice in the digital humanities have left me familiar enough with both sides to translate: the requirements of a funding application as readily as the limits of a data model that will later have to carry tens of thousands of records.

Wikibase and linked open data

One focus is Wikibase, the software behind Wikidata. Anyone running a Wikibase instance eventually faces the same questions: how do you model uncertainty? How does data get in at scale — and back out again? How do you keep it tied to authority files without starting over with every change?

The tools in the portfolio grew out of that practice: from conversion into the German national authority file, to reconciling entire archival holdings, to a visualisation that turns a SPARQL query into a diagram you can actually read.

Within the research community

As a member of the FactGrid Advisory Board I take part in decisions about the database that serves as infrastructure for many historical research projects in the German-speaking world. Alongside that, I work closely with projects from NFDI4Memory, the German National Research Data Infrastructure consortium for the historically oriented humanities.

This proximity to the research community is not a footnote but a working condition. Building tools for the humanities means knowing which standards apply, which repositories are eligible, and what funders will expect five years from now.

How I work

Open formats, open source, no lock-in: what I build should keep running even if you later choose someone else. That is why every project in the portfolio is publicly visible — you can read the code before you talk to me.

Funding periods end; research results are meant to last. That tension shapes every technical decision: as few moving parts as possible, as much as possible that works without anything running, and a clear route into a repository for whatever has to stay.