I build production-grade forecasting and optimization systems for critical infrastructure. Currently that means forecasting and trading in electricity markets. Before that: structural and climate-risk research at TNO in the Netherlands, and co-founding a humanitarian-tech startup that built satellite-based risk analytics for non-profits working in climate- and conflict-affected regions.

14+ years applying statistics and risk analysis across domains, 10+ years in machine learning with a focus on time-series forecasting, and 14+ years working on optimization methods. Python for 15+ years.

The combination of data science and natural-sciences training (a PhD in structural reliability, computational mechanics background) is what lets me formulate problems where the standard textbook formulation doesn't apply, not just fit a model to a clean dataset.

Moving between that many contexts trains you out of trusting consensus by default; you learn to go find the actual mechanism instead. It's also probably why colleagues have generally come to me for an unfiltered opinion on their work, not a polite one.

See Experience for the work history, Skills for the toolbox, or my LinkedIn profile.