Summary
Gary Zibrat is a seasoned software engineer with 12 years of experience specializing in GPU virtualization, operating systems, and AI/ML infrastructure. He has driven low-level platform work at Google on GPU VM families, contributed to kernel and microcode projects at Microsoft, and continued to shape infrastructure at Anthropic before joining OpenAI. Gary combines deep systems expertise with practical security and performance focus—experience that spans porting OS kernels to experimental architectures and addressing large-scale Spectre/Meltdown mitigations. Based in Los Angeles and trained at Cornell, he brings a research-minded approach to production problems and a track record of shipping infrastructure that powers large-scale AI workloads. An understated but valuable trait is his comfort operating at the intersection of hardware, firmware, and cloud-scale software, enabling cross-layer optimizations that many engineers overlook.
12 years of coding experience
10 years of employment as a software developer
Bachelor's degree, Computer Science, Bachelor's degree, Computer Science at Cornell University
English