Mike is interested in applying deep learning methods (particularly self-supervised, unsupervised, and generative learning methods) to cross-disciplinary problems in astrophysics, earth observation (EO), medical diagnosis and imagery, and other fields. His PhD work concentrates on building deep learning models that can efficiently mine and process very large scale (~petabyte!) astronomical datasets. He is also currently using deep learning to extract useful information from EO imagery data -- there is a surprising amount of overlap between EO and astronomy.