This section presents selected projects that demonstrate analytical, computational, creative, and problem-solving work. Some are larger case studies, while others are focused builds or explorations.
Full-stack WC 2026 prediction game built with Next.js 15 and Supabase. Users can predict match scores, group winners, and the full knockout bracket, compete on leaderboards, and create private rooms. The app is supported by a custom Python ML engine using RF + XGBoost, 68 features, Poisson score prediction, and full tournament simulation. Retrospective WC 2022 evaluation reached 54.7% outcome accuracy.
2D evolutionary simulation where agents learn to find food using genetic algorithms � selection, recombination, and mutation over generations. Demonstrates emergent learning from random vector-based "genetic" data.
Pygame-based piano learning app featuring an alternative keyboard layout optimised for learning. Highlights active scales in real time, marks in-scale vs. out-of-scale notes (A/B), and supports both keyboard and mouse input — designed to make music theory intuitive at the instrument.
Desktop app for documenting, calculating, and managing immunohistochemistry lab protocols. Includes protocol search by shared components, concentration calculators, and an auto-dated virtual lab notebook.