preptify - my startup, where i was the technical cofounder. it's an ai-powered mock interview assistant that tracks your performance and gives personalized feedback. i was responsible for building the whole product.

built with: react, node, firebase, aws, gpt, figma, deepgram, and stripe.

emily's archive - a blog for my girlfriend. built the whole cms for this from scratch. articles, images, the works.

built with: next.js, node, postgresql, firebase, and slate.

aggie prof rec - backend dev for this tamu prof-comparing site. when i joined, slashed query times from >30 seconds to milliseconds by building a massive database of tamu profs, courses, and stats.

built with: next.js, flask, selenium, cron, and postgresql.

alphazero tictactoe - replicated alphazero paper but for tictactoe because i dont have google researcher level compute (@google tryna change that???).

built with: tensorflow.

agentic ai assistant - needs a better name. proactive gpt based voice assistant that can handle multiple complex queries at once, roughly siri level latency.

built with: python, gpt, whisper, elevenlabs.

neural net from scratch - built a high-performance neural network library achieving 99% accuracy on MNIST. based off of micheal nielsen's book. no tensor libraries, just python.

built with: python.