J. William Murdock
Also known as Bill Murdock
J. William Murdock is a computer scientist who helped build IBM's Watson, whose team received the 2013 AAAI Feigenbaum Prize after the DeepQA question-answering system won its 2011 match on Jeopardy! After a 2001 Georgia Tech PhD in Ashok Goel's Design & Intelligence Laboratory and a postdoc at the United States Naval Research Laboratory, he joined IBM in 2003. From the project's 2006 feasibility study onward, he developed DeepQA components for typing answers and weighing evidence from passages, coauthored the 2010 DeepQA overview in AI Magazine, and guest-edited the 2012 "This is Watson" issue of the IBM Journal of Research and Development. Today he is a senior principal machine learning engineer at Red Hat, working on Llama Stack RAG capabilities and evaluation frameworks.
Timeline
-
2001
A Georgia Tech PhD
In 2001, he earns a PhD in computer science at Georgia Tech, where he is a member of Ashok Goel's Design & Intelligence Laboratory. His dissertation is titled "Self-Improvement through Self-Understanding: Model-Based Reflection for Agent Adaptation."
-
By 2003
Naval Research Laboratory and IBM
After his doctorate, he works as a postdoc with David Aha at the United States Naval Research Laboratory. He joins IBM in 2003; by 2010 he is a research staff member at the IBM T. J. Watson Research Center.
-
2006–2009
Question understanding for Jeopardy!
From the initial feasibility study in 2006, he works on IBM's Jeopardy! challenge and develops DeepQA components for typing answers and evaluating evidence from passages. By April 2009, his question-understanding work reaches 60–70% accuracy on Jeopardy-style questions, about twice the rate of other question-understanding systems.
-
2010–2011
The DeepQA paper and Watson's win
With David Ferrucci, Chris Welty and nine other coauthors, he writes "Building Watson: An Overview of the DeepQA Project" for AI Magazine in 2010. Watson wins its 2011 Jeopardy! match against Ken Jennings and Brad Rutter, and in February 2011 he returns to Georgia Tech to lecture on the system.
-
2012
Editing "This is Watson"
In 2012, he guest-edits "This is Watson," a special issue of the IBM Journal of Research and Development.
-
2013
An AAAI Feigenbaum Prize
Along with the rest of the IBM Watson team, he receives the 2013 AAAI Feigenbaum Prize. The award recognizes the team for demonstrating that a synthesis of AI techniques, including symbolic knowledge representation, natural language understanding and statistical machine learning, can achieve human-level performance in real-time factual question answering.
-
2018–2024
Watson Discovery and dialogue systems
By September 2018, he is a researcher at IBM's Watson Research Center and a development manager for the IBM Watson Discovery product. A 2022 NAACL-HLT industry-track paper he coauthors makes neural dense retrieval for dialogue systems run at scale on inexpensive hardware, and a 2024 paper fine-tunes a large language model into a zero-shot slot-filling system for conversational assistants.
-
As of October 2026
Llama Stack RAG at Red Hat
At Red Hat, he is a senior principal machine learning engineer who works on Llama Stack RAG capabilities and evaluation frameworks.
Sources and credits
Photo: We have no published photo of J. William Murdock yet. Send us a free-licensed photo.
- J. William Murdock - The Mathematics Genealogy Project
- Distinguished Lecture: Bill Murdock, Designer of IBM's Watson | Mercury
- The AI Behind Watson - The Technical Article - AAAI
- J William Murdock
- Murdock works with IBM's Watson – Design Intelligence Lab
- ISSIP Cognitive Systems Institute Group Speaker Series : September 13, 10:30am US Eastern Time - ISSIP
- AAAI Feigenbaum Prize - AAAI
- [2205.14226] Fast and Light-Weight Answer Text Retrieval in Dialogue Systems
- [2406.08848] An Approach to Build Zero-Shot Slot-Filling System for Industry-Grade Conversational Assistants


