Terrain Adaptive Robot Navigator
A 2WD chassis that classifies the surface underneath it, changing the way it drives — running entirely on the microcontroller.
I build full-stack applications and bring AI/ML into them — turning models into products people actually use
What drives me is building full-stack applications end to end — the database, the API, and the interface someone actually clicks through. The part I get most excited about is bringing AI/ML into that stack: training and fine-tuning my own models, then wiring them into real products so they solve problems instead of sitting in a notebook.
University of Waterloo Library
A 2WD chassis that classifies the surface underneath it, changing the way it drives — running entirely on the microcontroller.
Reads course material and generates practice that adapts to gauge a student's actual understanding, not just memorization. Built in 36 hours at the CxC AI Hackathon.
A digital accessibility tool generating AI-powered visual aids from lecture content, addressing a real unmet need for neurodivergent learners.
A Random Forest regression model trained on match history that predicts how many points a player will generate over their next 5 games.
A daily tech news digest for aspiring software engineers. Each day it gathers the most relevant stories, summarises them, and explains why they matter for a growing software career — with a simple todo list built in.