Senior engineer · energetic curiosity · hands-on builder · explorer

Portfolio

End-to-end AI systems wired to real applications — an LLM/RAG pipeline that answers questions against digitized archives, and a voice-agent workflow that turns phone calls into structured data.

RCA Archive Research

LLM/RAG pipeline for historical missionary archives. A LangGraph agent orchestrates retrieval, answer generation, and citation across digitized documents — ask questions in English or Chinese and get answers grounded in source passages.

RAG PipelineLangGraph AgentFastAPIPython

→ /rca-archive

VVOE (Verbal Verification Of Employment)

Voice-agent workflow: collect inputs, trigger an outbound call, and extract structured data from the conversation recap.

Voice AIWebhooksNode.js

→ /demo