Browser-based AI opponent with real-time gameplay, implementing Minimax with alpha-beta pruning in C++ and integrating it with a React frontend. Improved user interaction through adjustable difficulty levels via dynamic depth control and efficient game state evaluation.

Full-stack analytics platform for taxi driver revenue tracking, featuring SQL-based earnings aggregation, peak-hour profitability analysis, and an LLM-powered insights engine that generates automated weekly performance recommendations and structured tax reports.

End-to-end predictive analytics system training Random Forest and regression models on 60+ engineered features to predict NHL game outcomes. Includes reproducible Pandas/NumPy data pipelines and a CLI + RESTful Flask API for downstream inference, evaluated using cross-validation, F1-score, and RMSE.

Animated visualizations demonstrating GPU optimization concepts — memory coalescing, shared memory tiling, and roofline performance analysis. Built a modular rendering pipeline with batch processing and ffmpeg video export, programmatically illustrating warp-level memory transactions and matrix transpose optimizations.

Android event-registration app replacing first-come sign-ups with a fair lottery system. Supports entrant, organizer, and admin roles with QR code–based event discovery, randomized waitlist sampling with replacement, real-time notifications, image uploads, and optional geolocation verification.

Full-stack web application built at a University of Alberta hackathon to streamline discovering and applying for volunteer and job opportunities on campus. React.js frontend, Node.js + Express.js backend, SQL database — deployed to the cloud for high availability.

Real-time object detection system leveraging computer vision techniques to identify and track objects from live video streams with low-latency inference.