Rogier van der Geer

Field CTO at Xebia Data

I build AI systems that reach production rather than staying a demo. Most of my work is forecasting that carries real money, and agentic systems that have to hold up once real users get to them. Before consultancy I did a PhD in particle physics, which is where I learned to test a result before believing it.

Talks

  1. 2026From Data to Opportunity: Uncovering gaps in the market with Multi-Agent AIData Expo, Utrecht
  2. 2025Lesson Learned: How to get the most out of your agents?GoDataFest, Amsterdam
  3. 2024Scaling Up Albert Heijn’s AI-Powered Demand ForecastingEmerce EDAY, Amsterdam
  4. 2024Forecasting at ScaleFood & Data Symposium, ’s-Hertogenbosch
  5. 2023Streamlining Demand Forecasting at ScaleData Council, Amsterdam
  6. 2022Protocols in Python: Why You Need ThemEuroPython, Dublin
  7. 2019How to train an image classifier using PyTorchEuroPython, Basel
  8. 2017Playground of Evol AdventuresPyData, Warsaw
  9. 2017Risk AnalysisPyData, Amsterdam
  10. 2016How to conquer the worldEuroPython, Bilbao

Open source

pydargs
Configure a command-line application from a (Pydantic) dataclass.
evol
A grammar for evolutionary algorithms and heuristics. With Vincent D. Warmerdam. No longer maintained, but still fun to play with.

Elsewhere

Thesis

Thesis cover: Feynman diagrams drawn in rippled orange sand

Searches for New Physics Through Single Top
PhD thesis, University of Amsterdam and Nikhef, 2015. The work was done on the ATLAS experiment at CERN.