Ph.D. in Computer Science. ML/NLP scientist focused on information retrieval, conversational AI, and agentic systems. Builds production ML pipelines that improve answer quality and operational efficiency, from data and modeling to evaluation and deployment.

• Retrieval-augmented generation (RAG) and hybrid retrieval systems
• LLM/agent workflows, evaluation, and benchmarking
• Question answering, knowledge extraction, and explainable AI
• Production ML systems and data/annotation pipelines

Legal-RAG — https://github.com/Fan-Luo/Legal-RAG

• Statutory-focused legal intelligence system with verifiable citations and flexible LLM integration
• Multi-channel, graph-augmented retrieval pipeline (sparse, dense, late-interaction) with reranking
• FastAPI service and web UI with evidence panel for explainability and auditability
• Public demo via Hugging Face Space, Colab notebook, and video walkthrough

Multi-Agent Contract Platform — https://github.com/Fan-Luo/multi-agent-contract-platform

• Enterprise-grade contract comparison, redlining, and legal intelligence platform
• Clause-level semantic comparison (similarity + entailment/contradiction labels)
• Legal risk detection with structured findings, bias signals, and rationales
• DOCX redlining with tracked changes and in-document anchors
• Production-ready architecture (API, vector DB, LLM routing, agents, workers)
• Observability (structured logs, OpenTelemetry hooks) plus audit events

• Applied Scientist, Amazon Web Services (July 2022 - May 2025)
• Computational Language Understanding Lab, University of Arizona (2017 - 2022)
• Applied Scientist Intern, Amazon Web Services (June 2021 - August 2021)
• Insight Computer Architecture Lab, William and Mary (2015 - 2016)

• Ph.D. in Computer Science, University of Arizona
• M.S. in Computer Science, William and Mary
• B.S. in Computer Science, Beijing Institute of Technology

1. Fan Luo. Towards the Advancement of Open-Domain Textual Question Answering Methods. PhD Dissertation, 2022.
2. Fan Luo, Mihai Surdeanu. A STEP towards Interpretable Multi-Hop Reasoning: Bridge Phrase Identification and Query Expansion. LREC 2022.
3. Fan Luo, Ajay Nagesh, Rebecca Sharp, Mihai Surdeanu. Semi-Supervised Teacher-Student Architecture for Relation Extraction. NAACL 2019 Workshop.
4. Fan Luo, Mihai Surdeanu. Perturbation-based Active Learning for Question Answering. WNLP 2023.
5. Fan Luo, Marco A. Valenzuela-Escarcega, Gustave Hahn-Powell, Mihai Surdeanu. Scientific Discovery as Link Prediction in Influence and Citation Graphs. TextGraphs-12, NAACL 2018.
6. Fan Luo, Mihai Surdeanu. Divide & Conquer for Entailment-aware Multi-hop Evidence Retrieval. NAACL-HLT SRW 2022.
7. Rebecca Sharp, Adarsh Pyarelal, Fan Luo, Mihai Surdeanu et al. Eidos, INDRA, & Delphi: From free text to executable causal models. NAACL 2019.
8. Haonan Wang, Fan Luo, Mohamed Ibrahim, Onur Kayiran, Adwait Jog. Efficient and Fair Multiprogramming in GPUs via Effective Bandwidth Management. HPCA 2018.

• LinkedIn: https://www.linkedin.com/in/fan-luo-nlp
• Homepage: https://fanluo.me/