He does research in the intersection of computer vision and machine learning. More specifically, he is interested in self-supervised representation learning and adversarial robustness of deep models.