A (very) few of my past projects include:
- PoolShark
PoolShark is a reinforcement-learning AI pool player that learned to play 9-ball at a professional level (spin, position play, safeties, kicks). Built in 6 weeks using an AlphaGo-style approach: value/policy networks + Monte Carlo tree search, trained via self-play. PhD researchers at Cornell reached out after launch and we are now co-writing a paper.
- Companion AI
Companion is your personal AI tutor for learning to code - guiding you through exercises, side projects, and offering real-time help. It’s like having a tutor available 24/7 as you learn, practice, and build.
CreatorCoin serves as a medium to help individuals fund their personal side projects through NFTs.
- Natural Language to SQL Playground
A playground to convert natural language to SQL queries.
- Conversational AI Search over ArXiv Research Papers
A conversational AI search over ArXiv research papers.
- Exploring alternative AGI methods
Played around with Assembly Calculus, Ben Gortzel's work, evolutionary algorithms, and more.