I am a Machine Learning Researcher working on developing ML-based methods for downscaling climate models for the Antarctic, working on project ExtAnt.
I am particularly interested in extreme weather events and employing ML to effectively predict these events in a changing climate.
Research interests
I am interested in physics-informed ML and applying a scientific understanding to modern deep learning algorithms.
During my PhD at Helmholtz-Zentrum Hereon and the Max Planck Institute of Meteorology, I developed SuperdropNet — a machine learning based emulator of Lagrangian cloud microphysics to improve the representation of clouds and precipitation in climate models.
I also worked on developing strategies to create hybrid models that couple atmospheric models with machine learning components. We have coupled SuperdropNet to the ICON weather model and continue to optimize computational efficiency to make hybrid models a reality.
