I'm an AI/AX Workflow Automation Engineer focused on LLM Integration & Agentic Workflows — building MCP servers, RAG pipelines, and evaluation automation with Python/FastAPI (UC Irvine CE, on leave). I learn by shipping small, useful systems that connect real data to clear user workflows.
- AI/AX workflow automation — MCP servers, RAG pipelines, evaluation automation
- LLM Integration & Agentic Workflows for repetitive research, reporting, and operations
- Finance tooling — stock screening, brokerage dashboards, research packets (grounded generation)
- Local-first apps simple enough for non-developers to run
Building reliable systems first, then extending into LLM evaluation and RAG operations — the two new tracks below complement the existing data and API work.
| Project | What it is |
|---|---|
| K-invest | Read-only investment-data MCP server (Toss, KIS, SEC EDGAR, yfinance, Binance) for web LLMs |
| llm-eval-pipeline | Offline LLM evaluation harness: synthetic data → rule/proxy + optional LLM judge → error analysis |
| rag-ops-console | RAG operations sandbox: Chroma VectorDB + LangGraph + feedback monitoring |
| signal-foundry | Stock-screening engine that emits PDF research packets for people and MCP for agents |
| koreainv-dashboard | Personal KIS account dashboard across desktop, web, and Android |
| Project | What it is |
|---|---|
| zoomcheck | Zoom attendance dashboard that auto-matches and queues only ambiguous cases |
| margin-ta | Technical analysis engine for margin and trend signals |
| sat-qbank-exporter | Local-first exporter that turns the College Board question bank into printable PDF sets |
Python · JavaScript/TypeScript · C# · Kotlin · FastAPI · Electron · ASP.NET Core · Android