Gregory Diamos is a systems architect and startup builder with 12+ years advancing high-performance AI infrastructure, currently leading ScalarLM development to train LLMs with superintelligent reasoning capabilities. Based in the San Francisco Bay Area, he blends deep GPU architecture expertise—from inventing Volta’s independent thread scheduling to building LLM systems on AMD MI series hardware—with product instincts honed by founding and exiting an LLM startup. A co-founder of MLCommons and creator of industry-standard MLPerf benchmarks, he has repeatedly moved research into production at scale, including shipping large-scale models at Baidu and prototyping cutting-edge hardware solutions at AMD. Gregory advises multiple AI companies while holding senior roles focused on applying advanced reasoning models to relational data and real-world products. He holds a PhD from Georgia Tech and is known for combining low-level compiler and hardware work with leadership that turns benchmarking and systems research into widely adopted engineering standards. An unusual through-line in his career is his knack for bridging compiler/GPU innovations with practical large-scale LLM deployments, enabling both performance gains and broader industry adoption.
12 years of coding experience
10 years of employment as a software developer
Doctor of Philosophy (PhD) Electrical and Electronics Engineering, Doctor of Philosophy (PhD) Electrical and Electronics Engineering at Georgia Institute of Technology
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.