AI / ML Engineer
Documented
like a system.
I build RAG systems, reinforcement-learning agents, and production NLP. This site is the documentation: read the write-ups, play the game behind the paper, or ask my AI twin.
Built withHybrid RetrievalpgvectorGeminiWhisperXFastAPIPPO
RAG systems
grounded, cited, in production
RL agents
PPO, trained from scratch
Production automation
pipelines that still run daily
300+ students
ran exams on tooling I shipped
Published paper
on the Hunter Wumpus RL agent
Patent application
filed from the same work
At a glance
- Current
- AI & ML Associate, Impact Solutions Lab
- Education
- B.Tech, Artificial Intelligence & ML
- Projects
- Hunter Wumpus, AI Twin, and more
- Changelog
- What shipped, when
Featured work
PlayHunter WumpusA PPO agent hunts you through fog of war. Playable here, with a guided tutorial, and the subject of a published paper.QuestionAI TwinA retrieval-grounded chatbot that answers from a curated corpus of my work, with its retrieval trace on display.InspectSystem ArchitectureHow the site, the retrieval pipeline, and the game backend fit together, drawn out in diagrams.ReadThe PaperAdversarial reinforcement learning in a partially observable dungeon. The research behind the game.
The rest reads like docs.
Everything beyond this point is documentation, not marketing.