Junior actuary at Marsh Re, with
five years of models in production.
I work in analytics for reinsurance broking at Marsh Re in Munich. Before that I spent four years on bidding and forecasting models at Holidu, then a year on rate studies for US water and sewer utilities. I'm working through the DAV actuarial exams and work in English and German.
Four projects.
Two at Holidu, one in municipal finance, one from university.
Two models, scored against the market.
I built and ran both myself, from data pipeline to dashboard. Every forecast was logged before the game and scored after it. Both dashboards are still up with the full record.
World Cup 2026 model.
Forecast all 100 matches, each logged before kickoff. The model's log-loss was 1.003, better than an Elo baseline (1.010) and a coin flip (1.099). The market did better still (0.854–0.889) and won 64 of the 100 head-to-heads, though the knockout rounds were even. Team strength was built up from player ratings, each match and the full bracket were simulated 10,000 times, and an agent updated every forecast once lineups were confirmed.
March Madness model.
Tested out of sample on 11,170 games over two seasons: 71.6% accuracy, Brier 0.183, AUC 0.772. On the 4,240 games with market odds, the market was slightly better (71.8% and 0.183 against 70.2% and 0.192). The model is XGBoost trained on 16,691 Division I games, with walk-forward validation, calibration checks, SHAP explanations, player-availability adjustments and a 100,000-run bracket simulation.
How I work.
I'm trained in statistics and econometrics, and I build models that other people use to make decisions.
I pick the method to fit the problem. Usually that's a GLM; sometimes it's a bootstrap, a time-series model or gradient boosting.
Most of my work has been close to real money: ad spend, utility rates, and now reinsurance. I try to build models the person making the decision can follow, and to automate the routine work around them. At Holidu that roughly halved the team's weekly market-by-market checks.
I live in Munich, speak English and German, and am working through the DAV actuarial exams alongside my job at Marsh Re.
Experience & education.
The short version.