Gabriel studies the role of structure in deep learning models. Their current focus is on model design for improving CryoET reconstruction.  Their focus is on understanding how the structure of a deep learning model can be aligned with the physics of microscopy and biological structure for better and more reliable imaging.