My research lies at the intersection of causal learning and deep learning. Causal models generalize probabilistic models by adding the capability to infer cause-and-effect relationships among variables and events. Deep learning systems are increasingly being asked to answer questions that are essentially causal in nature. What is the causal effect of an agent's action on the state of the world and its ability to reach a goal? Which genes cause a disease? While deep learning is a powerful tool, it can only answer these questions reliably if it is used in a way that has a causal grounding. My goal is to exploit synergies between causal learning and deep learning, obtaining the inferential strength of causal mechanisms with the scalability and representational capabilities of deep learning.