Full-stack personal finance app built with React 19/TypeScript and Express.js/MongoDB. Features: transaction tracking, budget management with visualizations, savings pots, recurring bills, and a dashboard overview. Includes JWT authentication with Google OAuth, TanStack Query for state management, and a responsive UI with Radix UI components. Demonstrates RESTful API design, secure authentication, form validation with Zod, and modern React patterns including code splitting and error boundaries.

A climbing app where you snap a photo of any wall and instantly see optimal routes mapped out using OpenCV. Track your climbs, record attempts, and build your climbing portfolio with computer vision route detection. Send combines YOLO-based hold detection with intelligent route planning to help climbers visualize paths up any wall. The app includes session tracking, progress analytics, and real-time climb recording to help you improve your technique and reach your goals.

Implemented a custom shell environment that unifies standard command-line utilities with a graphical file browser. The core architecture is built around a custom command execution engine capable of parsing and managing both internal built-ins (like cd, ls) and external system processes.

This project connects real-time hand tracking on a computer to a physical servo-driven arm using a wired Arduino connection. A Python process captures video from a webcam, detects a single hand, derives finger bend angles, and streams those values over serial. An Arduino sketch receives the data and drives five servos in near real time.

Designed a Pro Micro-based mechanical keypad and developed a companion web-app and automation script to enable drag-and-drop firmware reprogramming without manual compilation.

Built an Arduino-based PID ventilation system prototype with C++. Used Python for performance data analysis to generate technical sustainability reports.

Developed a Verilog-based 2D engine in Quartus, implementing VGA generation, FSM game logic, and PS/2 input for real-time 60 Hz gameplay. Validated all timing and memory pipelines using ModelSim to ensure stable hardware synthesis.

A command-line game where users guess a randomly selected NBA player using structured attribute-based guesses, with clean input parsing and class-based game logic.

A terminal-based Reversi game in C where you play against an AI opponent. Supports configurable board sizes from 4x4 up to 26x26.