My research lies in the intersection of computer vision and machine learning. In particular, my research focuses in how we can leverage these to better understand and model complex human activities. I am interested in studying what are good representations for representing these activities, how to distinguish goals from preferences when humans are performing them, and how to learn good policies for agents to help in performing those activities. Besides this, I am also interested in the intersection of vision and language, the combination of neural and symbolic systems for computer vision, semantic segmentation and video understanding.