Summary
Paul Mcvay is a research scientist and machine learning practitioner with a PhD in Electrical Engineering from Texas A&M and nine years of industry experience spanning Waymo, Meta FAIR, and production ML at Overstock. His work blends theoretical guarantees—proved bounds on linear separability after compression and generalization for hard-SVM and multiclass coding—with practical systems, including a fast coding-matrix optimizer and productionized contextual bandits and Q-learning on GCP. As founder and sole developer of Winning Metrics, he built an end-to-end analytics platform used by 38 college programs, scraping and processing thousands of games and players annually. Equally comfortable in theory and engineering, he has scaled experiments to 200+ GB on AWS, implemented high-performance simulation code in Julia/C++, and driven cross-team ML infrastructure. Based in Morrison, Colorado, he seeks roles where rigorous theory meets production impact and continuous learning.
9 years of coding experience
10 years of employment as a software developer
Bachelor's degree Mechanical Engineering, Bachelor's degree Mechanical Engineering at Colorado School of Mines
Doctor of Philosophy - PhD Electrical and Electronics Engineering, Doctor of Philosophy - PhD Electrical and Electronics Engineering at Texas A&M University