I am a PhD student at Helmholtz Munich, supervised by Niki Kilbertus and Veronika Eyring. As part of the CWI PhD Internship Program, I had the pleasure of spending three months with the Scientific Computing Group in Amsterdam under the supervision of Daan Crommelin.

My research lies at the intersection of machine learning and Earth system science. I work on machine learning for weather forecasting and climate modeling, with a particular focus on uncertainty quantification, probabilistic machine learning, and the representation of unresolved processes. More broadly, I am interested in stochastic and data-driven model components, hybrid physical–machine learning approaches, and the long-term stability and generalization of machine learning methods in Earth system modeling.

My current work focuses on generative and autoregressive models for climate ensembles. The goal is to develop machine learning methods that can generate physically plausible ensemble members and thereby help quantify internal variability and uncertainty in climate projections, particularly in settings where large ensembles are unavailable.

Previously, I investigated stochastic machine learning parameterizations in chaotic dynamical systems, studying how different sources of uncertainty contribute to ensemble spread and forecast variability. Before that, I developed methods for training physically consistent and interpretable deep-learning-based climate model parameterizations by identifying and removing spurious non-physical relationships in the training data.