Projects

Projects

• Implemented a software based raycasting algorithm with vanilla JavaScript where parameters and the scene can be manipulated

• Solo submission made in 72 hours with vanilla JavaScript on an HTML canvas for the UW Game Dev Club's fall 2024 game jam; voted winner of the technical achievement award
• Built a real time optimized ray caster with ordered Bayer matrix dithering
• Programmed 2D rigid body collision behaviour and a randomized Prim's algorithm for map generation

• Team submission made in 72 hours with Godot for the UW Game Dev Club's spring 2024 game jam
• Implemented colour mixing, screen wrapping, movement, and flower spawning mechanics

• Visual file explorer of an IPFS written in Golang, Python, and JavaScript for the 2023 UofTHacks X hackathon; winner of the Estuary sponsor prize
• Created vector embeddings for files in the Estuary IPFS using Co:here’s NLP embeddings on file headers
• Performed principal component analysis of the vector embeddings to reduce dimensionality and be displayed interactively in 3D space with Three.js

• Several machine learning models written to predict the deterioration of the condition of the lungs over time for patients of Idiopathic Pulmonary Fibrosis (IPF) and give a level of confidence in the prediction
• Implemented an auto-encoder, linear regression, dense neural network, and bayesian model in order to accurately predict future lung capacity and give a confidence value using initial lung capacity data, age, sex, smoking status, and more
• Obtained a Laplace Log Likelihood score of −6.9 (much better than the baseline −8.1) with σ ≈ 200mL