After completing a bachelor’s degree in Electrical Power Engineering and Business Administration at RWTH Aachen University and interning at Tesla, I joined Chalmers University of Technology for the Data Science & AI master’s program in 2020. For my thesis project, I became part of the team at the healthy AI lab and was supervised by Fredrik Johansson. Since then, I have joined BCG Gamma as a data scientist.
During my master studies I participated in a research project with E.ON and IAEW of RWTH Aachen (https://www.iaew.rwth-aachen.de/go/id/cyffs/?lidx=1), where we studied how one can use graph neural networks and reinforcement learning to make smart decisions in energy distribution grids in order to deal with the intermittency of renewable energy sources. For my master thesis project I studied how one can increase the sample efficiency of nonlinear machine learning models by incorporating privileged information, which is data that is only available at training time but not when a model is finally used. For the setting where the prediction target is related to the model input through the privileged information in terms of a time series, we showed that privileged learning is beneficial under certain assumptions. In addition we introduced privileged learning algorithms that show performance benefits over classical learning empirically. Settings where privileged information is available are common in many applications, especially in healthcare, meaning one can hope to make more accurate predictions from small data sets using these techniques. The paper about this work can be found here (https://arxiv.org/abs/2209.07067).