I'm a full-stack developer and graduate student at Chengdu University of Technology, working on AI SaaS products, LLM agent systems, RAG applications and data-driven backend platforms.
My recent work focuses on turning AI capabilities into reliable product systems: routing, streaming, permissions, billing, retrieval, observability and frontend workflows.
FindRealTaste Repo · Live Preview
- Built dynamic AI provider routing with Bayesian scoring, P95 latency normalization and circuit breaker isolation.
- Designed Redis Stream + SSE event pipelines for long-running agent tasks with reconnect recovery.
- Implemented multi-tenant permission models covering API authorization, row-level scope and field-level masking.
- Built hybrid retrieval pipelines with vector search, full-text search, trigram matching and RRF rank fusion.
- Developed AI-assisted workflows with Claude Code, Codex, Cursor and GitHub Copilot to improve delivery speed and code quality.
- Experienced with Vue ecosystem development, frontend admin systems and map/GIS interaction scenarios.
I'm interested in practical AI products: systems that combine LLM reasoning with strong engineering foundations, including permissions, billing, streaming, retrieval, evaluation and production reliability.
