Data science grounded in statistical inference — an economist's training in econometrics and Bayesian methods, applied to machine learning problems end to end.
Bachelor thesis (University of Cologne): nowcasting US recession probability by combining macroeconomic, financial, and text-based predictors, estimated under both Bayesian and frequentist frameworks — a direct comparison of the two inferential paradigms on the same forecasting problem.
Write-up available on request
A full pipeline on real, messy data: scraping player performance and market data from fbref and Transfermarkt, feature engineering, gradient-boosted regression with randomized hyperparameter search, and feature-importance analysis of what actually drives transfer valuations.
A variational autoencoder built as structured software rather than a notebook — CLI, data loaders, and loss modules — generating MNIST digits in grayscale and color. Hands-on variational inference: KL divergence, the reparameterization trick, and learned latent spaces. Graded A.
Product-department internship at a B2B fintech (buy-now-pay-later for business trade).
Built Python tools (Pandas, NumPy, Matplotlib, Seaborn) to analyze R&D measurement data; automated hazardous-material labeling with an Excel/VBA tool highlighted by the employer as a standout achievement; developed automation tooling for a department digitalization initiative and collected measurement data used to train AI models.
Self-directed specialization in quantitative methods: econometrics, multivariate data analysis, empirical methods, and market design. Thesis supervised at the Institute of Econometrics and Statistics.
I'm looking for a graduate data science role starting mid-2027 — Oslo, Germany, or elsewhere in Europe. If rigorous inference on real business problems is what your team does, I'd like to hear from you.