I am currently a research scientist at Meta AI (FAIR), an adjunct professor at McGill University, a core industry member of Mila, and a Canada CIFAR AI Chair. The playground of my research has been defined by problems which require inferring full observations from partial observations, building models of the world with the goal to improve impactful downstream applications. My most recent research explores the potential of conditional visual generative models as synthetic training data sources for downstream model training and self improvement. In particular, I am interested in the scaling laws of synthetic data, model sampling techniques to tailor synthetic data to the needs of the learner and to address model collapse. Prior to joining Meta AI/McGill/Mila, I received my Ph.D. from University of Barcelona, where I worked with Dr. Carlo Gatta, and I spent two years as post-doctoral researcher at Mila working with Prof. Yoshua Bengio.