Michael Ferreira

• —Led end-to-end deployment of Stripe payment processing, authentication, and user management across a React web app and iOS Swift app, supporting 100+ users and 1,500+ transactions
• —Designed and optimized a PostgreSQL database on Supabase powering 50+ API endpoints, achieving sub-100ms response times through caching, denormalization, and memoization
• —Delivered 10,000+ lines of production code across the full stack and mobile app, driving $5,000+ in revenue and enabling 5x user growth through feature development and performance improvements

• —Refactored the website to leverage SSR and SSG techniques, reducing client-side bundle size by 68% and improving initial page load time from 2.8s to 0.9s.
• —Developed YOLO computer vision pipeline for automated TV display recognition in retail stores, achieving 30+ FPS throughput with sub-500ms latency through OpenCV preprocessing; technology acquired by Taiv (YC W20)
• —Engineered a neural network audio classifier using MFCC, mel-spectrogram, chroma, and tonnetz features using librosa, achieving 92% accuracy with a 3-layer dense architecture optimized through 64-epoch training

• —Built and deployed a full-stack MERN donation platform serving 10 charitable organizations with 90+ active users, contributing 9K+ lines of production code that facilitated $1,000+ in donations
• —Implemented JWT-based authentication with bcrypt password hashing and RBAC protecting 50+ API endpoints while maintaining sub-100ms average response time
• —Optimized MongoDB schema with compound indexes on 3 core collections achieving 40% faster query performance and supporting 10x data volume scalability through denormalization strategies

A web application that connects students looking for sublets with those offering them, featuring search, listings, and user authentication.

A machine learning pipeline for predicting forex price movements using technical indicators and market data. Achieved significant accuracy in forecasting currency trends.

A platform designed to make charitable giving simple, transparent, and impactful by directly connecting donors and charities.

A replica of the classic Atari Centipede game developed using Python and Pygame as the final culminating project for my grade 12 computer science class.

A Python-based stock trading simulation game built with Pygame; based on the Stock Ticker board game where players buy and sell stocks to maximize their portfolio value.