Built a system that streams IMU data through the OpenAI API and displays bench-press feedback on rep count, imbalance, tempo and rest quality — on-device, in real time, on an ESP32.
Tested the ROME and MEMIT model-editing algorithms across GPT-2 models for targeted fact correction without full retraining — the core question behind privacy and the 'right to be forgotten.' The smallest model corrupted unrelated facts and the largest failed to hold edits under rephrasing; mid-size models performed best.
A Database Systems build with a five-person team. I owned the seller side - inventory CRUD, order fulfillment, stock - plus the UI/UX. Flask + PostgreSQL with complex queries; search, carts, checkout end to end.
A convolutional neural network that sorts waste imagery into recyclables, organics, and landfill - trained for sustainable waste-management workflows.
A Flutter + Firebase to-do app with authentication and real-time sync, a playful Dark Side aesthetic, and Android + Web support. Team project.
Monochrome, keyboard-navigable, statically exported to Cloudflare Pages at ~25KB per route. Next.js 16 + TypeScript + Tailwind v4, consent-gated Open-Meteo weather, a strict CSP, and a main/prod branch strategy with semantic versioning.