Artur Back de Luca

I’m a fourth-year Ph.D. student in Computer Science at the University of Waterloo, advised by Kimon Fountoulakis. My research focuses on reasoning in neural networks, particularly on which problems they can solve, how they can learn to solve them from data, and how we can certify that they have learned to do so correctly.

Over the years, I have also worked on other topics, including machine learning on graphs. At Amazon, I worked on probabilistic models for forecasting inventory arrivals across the supply chain and, more recently, on diffusion models to generate time-series data for training forecasting foundation models. Earlier, at Huawei’s Noah’s Ark Lab, I worked on federated learning through connections to domain generalization.

I expect to graduate in 2027. Feel free to reach out about industry opportunities where you think there might be a fit.