data · AI-assisted analytics · governed AI

I build governed AI analytics: the user steers, the AI stays at the border, and the analytical core stays auditable.

Explore how the work connects →

Welcome! I'm Hélène. I build analytics you can reproduce, defend, and audit, which means I take it slow. I don't give an answer in two minutes: I confirm the premises, show partial results, and steer before the full analysis runs, then shape the writing-up. Every step where judgment matters is confirmed and logged.

Most of my current work uses AI on research data in three ways: to augment the analysis itself, to enrich data, and to build the workflows that connect them: always governed, always auditable. The most architecturally involved of these is what I call an AI metascientist: a governed architecture, built so interpretation, execution, and validation stay separate. See how it all fits together →

My career started in geomatics nearly twenty years ago: spatial data, cartography, and the practical problem of representing complex data without lying about them. That is where my technical knowledge started; the data management, the visualisation, and the coding across both workflow and web. A research PhD gave me the research, the analytics, and the thinking that sits behind them. The objects I work with now are research organisations, publications, classifications, and funding systems, but the underlying question is unchanged: how far the data can take us, what it cannot support, and how to make that distinction clear enough for people to use. CV →