Will Boyd

Will Boyd focuses on post-training of foundation models and their applications in Microsoft’s Applied Sciences Group (ASG).  His research interests include foundation model post-training, reinforcement learning, and scalable machine-learning systems.  He is especially interested in the challenges of making advanced A.I. models more capable, reliable, and useful in product settings.

Before joining ASG, Will held research and engineering roles at Apple, Microsoft, MITRE, and MIT, where he helped turn emerging research ideas into product-oriented A.I. systems.  His recent work has centered on foundation model post-training, spanning infrastructure for scalable model adaptation as well as training models for specific applications.  Earlier in his career, he led and contributed to projects in visual understanding, human-centered interaction, and geospatial analysis, applying machine learning across consumer technology, cloud A.I. services, and national security research.

Will received his Ph.D. and M.S. in nuclear science and engineering from MIT, where his research applied high-performance computing, numerical methods, and machine-learning-inspired techniques to nuclear reactor simulation.  He was supported by the NSF Graduate Research Fellowship and the NRC Nuclear Education Graduate Fellowship.  During graduate school, he co-created and contributed to open-source neutron transport simulation codes, including OpenMOC and OpenMC.  Will also holds bachelor’s degrees in computer science and physics from Georgia Tech.

In his free time, Will enjoys spending time with his family, traveling, running, cycling, hiking, and backpacking.  He has thru-hiked the Appalachian Trail and section-hiked most of the Pacific Crest Trail.