Student developer focused on practical AI systems, LLM agents, and open-source developer tools.
I build small, useful systems for testing model behavior in real workflows: agent simulations, search-augmented APIs, and local-first analytics for AI coding tools.
LLM agents, model evaluation, model-provider integrations, search-augmented generation, developer automation, AI-assisted code review, and local-first tooling for privacy and cost visibility.
MIT-licensed AI trading simulator with 900+ stars and 270+ forks. It supports local and online play, OpenAI-compatible providers, realistic trading fees, leaderboards, and local data storage for privacy-aware experimentation.
A lightweight API enhancer that adds live web search to LLM APIs, helping model responses stay fresher and better grounded.
A local-first desktop app for tracking token usage and AI costs from Claude Code, Codex CLI, and OpenCode logs.
I am exploring maintainable LLM agent workflows, better model evaluation loops, and tools that help open-source maintainers review, test, and ship software with more confidence.