Chadwick Petey

Chad works on low-level optimization and model inference performance at Drane Labs. He's the team's go-to person for anything involving CUDA kernels, memory layout, or squeezing an extra ten percent out of a training run that everyone else thought was already optimized.

Before Drane Labs, he worked at NVIDIA on their cuDNN library team, where he contributed to several of the fused-operation kernels that are now standard in deep learning frameworks. He holds an MS in Computer Science from the University of Washington.

Outside of work, Chad is an avid pinball enthusiast — he owns three restored machines and has competed in regional tournaments. He's also a devoted retro gaming collector with a particular love for NES classics; coworkers say his all-time favorite is Contra.