I study Engineering Science at the University of Toronto. Right now I’m building an LLM inference gateway in Go. It sits between apps and AI providers like Cohere, and handles routing, rate limiting, caching, and failover.
Before that, I worked on caching at Shopify, and trained neural networks at UTIAS to keep drones stable while they carry a swinging payload. My team’s autonomous robot also took 2nd place at UTRA Hacks 2026.
Outside of that, I play oboe and piano, a few video games, and go hiking out West.
Work
What I’m building now, my internship at Shopify, research at UTIAS, and a few hackathons.
All projects on GitHubA gateway in Go that sits between apps and AI providers like Cohere. It routes requests, rate-limits them, caches responses, and fails over when a provider goes down. I’m working on how it schedules requests and handles concurrency, to bring latency down and throughput up.
Go
Footage: Shopify
On the Metafields team, I cut repeated callback work, moved two loaders onto one cache reader, and replaced 32 copies of capability logic with a GraphQL API.
Ruby · Memcached · GraphQL
Trained neural contraction metrics (neural CCMs) for stable trajectory tracking of drones carrying slung payloads, and extended them toward 360° roll maneuvers.
PyTorch · Neural networks · Control theory
UTRA Hacks 2026 · 2nd place overall · Best track course performance
An autonomous robot for a Winter Olympics–themed course, built with four teammates. It picks up a cube, drives either an obstacle course or a set of ramps, lines up on a target by colour and launches a ball, all from a modular C++ codebase on an Arduino.
C++ · Arduino · PlatformIO
An app that watches heart rate from a Fitbit and alerts people nearby if something looks like an emergency. Built at Hack the 6ix.
React · Node.js · Fitbit API · Twilio
A web app that predicts where lost items and oil spills drift at sea, using ocean-current data for East Asian waters. Built at Hack3.
Python · Flask · JavaScript