Biography
I am a Ph.D. student in computer science at the University of Chicago advised by
Hank Hoffmann. My research interests lie at the intersection of embedded systems, machine learning (ML), and computer security. We generally design ML applications that fit within the tight constraints of low-power devices. To accomplish this goal, we create ML systems that support adaptive behavior based on runtime feedback; this adaptivity allows the inference application to alter its execution based on resource availability. This adaptive behavior, however, comes at a cost in security. Adaptive systems change their functionality according to patterns in the input data. This behavior creates potential side channels: adversaries who observe the adaptive decisions (e.g. through device power or communication patterns) learn information about the collected inputs. Our goal is to design low-power adaptive systems that avoid leaking information through side channels on embedded devices.
Before starting at the University of Chicago, I did my undergraduate studies in computer science at the University of California at Berkeley. I then completed my master's (MPhil) in computer science at the University of Cambridge (Pembroke College). My master's thesis developed a method to train graph neural networks to solve flow problems using self-supervision derived from Lagrangian Duality.
