Postdoctoral Researcher
Dr. rer. nat. · Statistics & Machine Learning
TU Dortmund & Lamarr Institute for ML and AI
I develop scalable Bayesian methods and coreset theory for large-scale statistical inference, with applications spanning particle physics simulations and high-dimensional generative models.
I am a Postdoctoral Researcher at TU Dortmund University and the Lamarr Institute for Machine Learning and AI, one of Germany's leading AI research centers. I completed my PhD in Statistics in February 2026 with a dissertation on large-scale data reduction based on coresets.
My work sits at the intersection of theoretical statistics and practical scalability: I develop coreset methods that compress massive datasets while provably preserving the statistical properties needed for Bayesian inference. Beyond methodology, I apply these tools to challenging domains including particle physics (CERN/ATLAS) and multivariate generative models.
I am currently exploring opportunities in quantitative research, applied statistics, and ML research roles in industry — particularly in pharmaceutical statistics, tech, and finance.
Advances in Data Analysis and Classification, 2024
Machine Learning under Resource Constraints, deGruyter, Berlin, 58–70
TU Dortmund & Lamarr Institute for Machine Learning and AI
Continuing research on scalable Bayesian methods, coreset theory, and applications to particle physics simulation in collaboration with CERN/ATLAS.
TU Dortmund — SFB 876-C4 Research Group
Developed novel algorithms for large-scale Bayesian inference. Taught Monte Carlo Simulation, Bayesian Statistics, Statistical Learning & Big Data (100+ students per semester). Developed R package BayesPprobit (CRAN) and Julia package MCBench.
Daimler Mobility AG, Global Headquarters
Led AI for Credit Scoring project — improved global corporate rating model ROC from 77% to 86% using XGBoost, LightGBM, and Random Forest. Built automated reporting tools in VBA and SAS.
China Construction Bank, Frankfurt Branch
Developed R scripts for automated VaR calculation, backtested PD and LGD models, and prepared sovereign risk & liquidity stress-testing reports.
R Package · CRAN
Scalable Bayesian estimation for p-generalized probit and logistic regression models via coreset-accelerated MCMC. Handles large-scale binary classification problems with theoretical guarantees.
Julia Package
A comprehensive benchmark suite for Monte Carlo sampling algorithms. Provides standardized test distributions, convergence diagnostics, and performance metrics for comparing MCMC methods.
Collaborative Research · TU Dortmund + CERN
Developing Bayesian and Monte Carlo algorithms for rapid simulation of particle collision experiments using ATLAS data. Applying GANs, normalizing flows, and diffusion models.
Research · TU Dortmund
Novel coreset compression algorithms for multivariate conditional transformation models. Theoretical proofs of approximation guarantees with empirical validation across simulation studies.
Open to research collaborations, industry opportunities, and academic exchanges.
