Quan
Bui
Applied AI Engineer · Forward Deployed
I ship production LLM systems — RAG, agents, evals — and work directly with customers to get them there: POCs, solution architecture, and technical direction. Six years of NLP research underneath it all.
Agentic RAG · Azure
Agentic RAG Chatbot for University Student Services
A prior vendor's RAG deployment failed in production — no metadata, no reranking, broken document processing. I rebuilt it as a chatbot agent on Azure with parent-child retrieval and clarification tools, serving 40,000+ students and staff.
Trade-off: patch the existing RAG pipeline vs. rebuild as an agent → rebuilt as an agent so it could ask clarifying questions and filter by metadata before retrieving, cutting peak-enrollment tickets ~50%.
Dr. Quan Bui
PhD · JAIST · Japan
I'm an applied AI engineer who ships production LLM systems and works directly with the people who use them. Day to day: pre-sales technical discovery and proof-of-concept (POC) builds, advising customers on solution architecture, and negotiating technical direction with stakeholders — then building the system and standing behind it in production.
Six years of doctoral research at JAIST in natural language processing and legal AI is the foundation that makes me a sharper builder and advisor — four-time COLIEE champion, ten-plus papers, still publishing. Research is how I learned to design evaluation frameworks, reason about agent architecture, and tell what actually works from what only demos well. It's why my POCs hold up when they meet real data.
That legal-NLP background is also domain expertise these companies sell into: regulated industries — legal, financial services, compliance — where LLM-in-production deployments live or die on accuracy, traceability, and trust. I've spent my career inside exactly those constraints.
And I write about what I build — 61 technical articles on context engineering, agent observability, RAG architecture, and harness engineering. Teaching the work is how I pressure-test it.