In my capacity as a scientist and academic, I study the design, architecture and algorithms of scalable and robust end-to-end ML systems. My research interests drive towards the development of new forms of ML systems that are revolutionary in how they leverage multi-modal data (e.g., speech, vision, location, inertial information) to infer and reason over complex real-world situations — while simultaneously, maintaining extreme levels of systems flexibility (e.g., distributed execution, adaptation) and efficiency (e.g., compute, memory). In the past, I have developed computational models of human behavior and context (i.e., activity recognition), in addition to algorithms and systems software for mobile and embedded platforms ranging from smartphones to sensor networks.