I'm a full-stack developer based in Ottawa. I've been building and shipping software for a couple years now, across web, mobile, ML, and systems programming. I work primarily in Python, JavaScript/TypeScript, Java, and Rust, and I'm comfortable across the full stack: React, React Native, FastAPI, Node.js, Express, PostgreSQL, Docker, and CI/CD with GitHub Actions. Projects ferrodb (in progress) is a relational database I'm building from scratch in Rust. Page-based storage, B+ trees, buffer pool management, write-ahead logging, MVCC, a cost-based query optimizer, SQL parsing, and Postgres wire protocol compatibility. This is a summer-long systems project focused on understanding how databases actually work under the hood. ForgeFind is an image forgery detection tool I built with a PyTorch U-Net model (92% accuracy on CASIA v2.0) and OpenCV SIFT for copy-move detection. It's served 1,000+ users. I optimized the Docker image by 62% (2.4GB to 910MB) and set up automated deploys through GitHub Actions. Live at forgefind.netlify.app. pic2recipe is a mobile app (React Native/Expo) that identifies ingredients from food photos using the Gemini API and matches them against 2.2M recipes via pgvector similarity search. I brought query times from 8s down to 150ms with ANN indexing and cut peak RAM from 6.4GB to 50MB using numpy memmap. Full backend in FastAPI and PostgreSQL, containerized with Docker. DeadWeight is a website carbon waste scanner built with Next.js, Express, and Puppeteer. It identifies unused code and removable bloat across multi-page scans. Won 1st place out of 30 teams at the Hack Club Atlas Hackathon. Other experience I spent a year as a software developer on FRC Team 7476 (FIRST Robotics Canada), writing Java control code for 3 robot subsystems using WPILib. I reduced response latency by 33% through PID tuning and helped debug live control and sensor issues across 4 competitions. The team qualified for the FIRST World Championships. I also tutor CS students at Auxilium, where average test scores have improved by 15%. Tech I use: Python, Java, Rust, JavaScript, TypeScript, SQL, React, React Native, Expo, Node.js, Express, FastAPI, PyTorch, OpenCV, Docker, GitHub Actions, PostgreSQL, pgvector, MongoDB, Linux.…I'm a full-stack developer based in Ottawa. I've been building and shipping software for a couple years now, across web, mobile, ML, and systems programming. I work primarily in Python, JavaScript/TypeScript, Java, and Rust, and I'm comfortable across the full stack: React, React Native, FastAPI, Node.js, Express, PostgreSQL, Docker, and CI/CD with GitHub Actions. Projects ferrodb (in progress) is a relational database I'm building from scratch in Rust. Page-based storage, B+ trees, buffer pool management, write-ahead logging, MVCC, a cost-based query optimizer, SQL parsing, and Postgres wire protocol compatibility. This is a summer-long systems project focused on understanding how databases actually work under the hood. ForgeFind is an image forgery detection tool I built with a PyTorch U-Net model (92% accuracy on CASIA v2.0) and OpenCV SIFT for copy-move detection. It's served 1,000+ users. I optimized the Docker image by 62% (2.4GB to 910MB) and set up automated deploys through GitHub Actions. Live at forgefind.netlify.app. pic2recipe is a mobile app (React Native/Expo) that identifies ingredients from food photos using the Gemini API and matches them against 2.2M recipes via pgvector similarity search. I brought query times from 8s down to 150ms with ANN indexing and cut peak RAM from 6.4GB to 50MB using numpy memmap. Full backend in FastAPI and PostgreSQL, containerized with Docker. DeadWeight is a website carbon waste scanner built with Next.js, Express, and Puppeteer. It identifies unused code and removable bloat across multi-page scans. Won 1st place out of 30 teams at the Hack Club Atlas Hackathon. Other experience I spent a year as a software developer on FRC Team 7476 (FIRST Robotics Canada), writing Java control code for 3 robot subsystems using WPILib. I reduced response latency by 33% through PID tuning and helped debug live control and sensor issues across 4 competitions. The team qualified for the FIRST World Championships. I also tutor CS students at Auxilium, where average test scores have improved by 15%. Tech I use: Python, Java, Rust, JavaScript, TypeScript, SQL, React, React Native, Expo, Node.js, Express, FastAPI, PyTorch, OpenCV, Docker, GitHub Actions, PostgreSQL, pgvector, MongoDB, Linux.WWWW…