I'm a Machine Learning & Data Science researcher and research software developer at the German Sport University Cologne. My work focuses on probabilistic modeling, end-to-end machine learning, and analytical software for complex real-world data.
- Machine Learning & Data Science: probabilistic classification, gradient boosting, Bayesian modeling, feature engineering, calibration, evaluation, and interpretation
- Research Software: Python, JavaScript/Vue.js, Git, Docker, and web-based analytical tools
- Applied Analytics: translating research questions and real-world problems into robust models and usable decision-support workflows
Beyond Outcome Bias
A probabilistic machine-learning framework for risk-aware soccer action evaluation. It introduces xSuccess to model contextual action-completion probability and integrates it into an adjusted VAEP formulation to reduce outcome bias.
xImpact
A multi-level probabilistic framework that links action-level evaluation to match-level objectives using xSuccess, risk-adjusted VAEP, and Bayesian in-game win probabilities.
SportVid
A DFG-funded platform for AI-supported video and data analysis in sport. I lead frontend development and contribute to technical integrations, infrastructure, and deployment-related work.
- Expected Impact on Match Outcome: A multi-level probabilistic framework for context-sensitive soccer action evaluation
- Beyond Outcome Bias: Incorporating action completion probability and risk-return into soccer evaluation models — MLSA 2025 / Springer, 2026
- The Importance of Positional Placement: Investigating the influence of different infield positions of floater players in small-sided soccer games — International Journal of Sports Science & Coaching, 2025
Python · pandas · NumPy · scikit-learn · XGBoost · PyMC · JavaScript · Vue.js · Docker · Git