We investigate a variety of open problems that sit at the intersection of artificial intelligence, machine learning, and computer systems. We investigate the development of novel algorithmic and theoretically principled methods that are grounded in mathematics for systems problems with the ultimate goal of building reliable and high-performance machine learning systems. In particular, we work on Causal AI, Transfer Learning, Multi-Objective Optimization, and Disentangled Representation Learning from the theory side. On the application side, we aim to develop the next generation of autonomous systems (on-device, embedded, heterogeneous, cloud, robotics) that can perceive, reason, and react to complex real-world environments and users with high levels of precision and efficiency. For the CIFellow proposal, I am interested in mentoring topics related to Causal AI, Representation Learning, Transfer Learning, and applications of these in Autonomous Systems.