Biography
I’m a Computer Science PhD student at the University of Waterloo, advised by Khuzaima Daudjee and supported by the NSERC CGS-D scholarship.
My research focuses on improving scalability and efficiency of LLM training in cloud environments. I’ve also worked on LLM inference, computer vision, and reinforcement learning, and I’m broadly interested in building large scale systems for the next generation of AI workloads. During my PhD, I’ve completed research internships at Microsoft Research, Apple, and Waabi.
I earned my Bachelor’s degree in Computer Science from the University of Waterloo in 2019. As an undergraduate, I interned at Uber ATG, Jane Street, LinkedIn, Yahoo, and OMERS Capital Markets.
Interests
- ML Systems
- Distributed Systems
- Cloud Computing
- Reinforcement Learning
Education
-
PhD in Computer Science, 2019 - 2025
University of Waterloo
-
Bachelor of Computer Science, 2015 - 2019
University of Waterloo
Publications
Sampling-based Predictive Database Buffer Management
Cephalo: Harnessing Heterogeneous GPU Clusters for Training Transformer Models
Rethinking Closed-Loop Training for Autonomous Driving
Hydrozoa: Dynamic Hybrid-Parallel DNN Training on Serverless Containers
Experience
Research Intern, Gray Systems Lab
Microsoft
• Developed and patented a scheduling algorithm, achieving up to 40% reduction in runtime in simulations
ML Intern, LLM Training
Apple
• Worked with internal client to optimize 13B LLM pre-training, achieving a 2x speedup on a 512-GPU cluster
Research Intern, Closed-Loop Training
Waabi
• Work was patented and published at ECCV 2022