About
I'm Andrew Lee. I build LLM products end to end: scoping the problem with stakeholders, designing the system, running the evals, and shipping it to real users. I'm based in Taiwan and work with teams internationally.
Highlights
- REPLACE: one quantified result, e.g. "Cut hallucination rate on a customer-support RAG system from 14% to 3% by rebuilding the retrieval and eval pipeline."
- REPLACE: one quantified result about scale, cost, or latency.
- REPLACE: one result about people or process, e.g. "Led a 6-person team through the first production deployment of an agentic workflow."
- REPLACE: optional fourth or fifth item.
What I'm working on
REPLACE: 2–3 sentences on the problems you are focused on right now (e.g. evaluation for agents, cost/quality trade-offs, how teams adopt AI tooling). Update this every quarter.
Working together
I'm open to REPLACE: full-time roles / advisory / speaking / collaboration on specific problems. The fastest way to reach me is email:hello@andrewlee.dev.
Elsewhere
I post in English on LinkedIn and in Traditional Chinese on Threads. Longer pieces always land here first.
REPLACE with one or two sentences of life outside work (where you live, one hobby, one thing you are learning). This is the icebreaker, not a diary.