Stefan Schroedl is an applied scientist at Alexa AI, where he is developing interpretability methods for model improvement and training data reduction. In previous roles, he led a machine learning team at Atomwise in building deep neural models for drug discovery; was responsible for relevance ranking for Amazon product search; worked on search advertising technologies at Yahoo, and on high precision maps at MercedesBenz R&D. Stefan holds a Ph.D. in computer science from the University of Freiburg, Germany.