I enjoy turning ideas into working products, from the interface and backend to the AI systems behind them. Most of my projects start with curiosity: build a small version, test it with real inputs, and keep improving the parts that break.
- AI agents — tool use, MCP integrations, workflow orchestration, memory, evaluation, and failure recovery
- Retrieval systems — RAG pipelines, query rewriting, reranking, source grounding, and retrieval-quality experiments
- Full-stack products — responsive web interfaces, APIs, databases, authentication, deployment, and observability
- Developer tools — practical ways to make AI-assisted development more reliable and reproducible
TypeScript · React · Next.js · Java · Spring Boot · Python · PostgreSQL · Docker · GitHub Actions
Lately, I have been especially interested in the boundary between a convincing Agent demo and a dependable system: what happens when retrieval misses, a tool fails, context grows, or an external service becomes unavailable.
If you are exploring similar ideas, feel free to look around or start a conversation through GitHub.