About

I am completing a PhD at McGill University’s School of Computer Science and I am affiliated with Mila. My supervisor is Christophe Dubach. I design compiler systems that make advanced program optimizations practical, with a focus on parallelism, hardware efficiency, and extensible design.

My research interests span optimizing compilers, equality saturation and emerging hardware. The intersection of these interests is the topic I pursue in the context of my PhD. Specifically, I am researching the applications of equality saturation to solve code generation challenges for Machine Learning (ML) accelerators and Field-Programmable Gate Arrays (FPGAs).

ML accelerators and FPGAs have tremendous compute power but are held back by unique programmability challenges that require dedicated solutions. The solutions I propose in my work include latent idiom recognition for ML acceleration libraries and automatic partitioning and resource sharing for FPGA designs. These solutions rely on functional program representations and equality saturation, a novel and increasingly popular technique in compilers.

Due to its relative recency, the broader field of equality saturation features a range of open problems, notably scalability and Intermediate Representation (IR) design. IR design is a common thread throughout my work. The scability challenge is also a topic I am exploring in my research.

My work has resulted in multiple peer-reviewed publications at top compiler venues (CGO, CC) and in open-source compiler infrastructure. I am currently exploring Canadian research and engineering roles in compilers, programming languages, and systems.