Caomaolufei (草帽路飞)

Hi, I'm Caomaolufei. I work on AI Infra for a foundation model team at a large internet company and have several years of experience in the field, mostly on model training, inference and kernel optimization.

I also wrote AIInfraGuide, an open-source tutorial that starts from zero and covers the whole AI Infra stack. It takes you from low-level CUDA programming through distributed training and inference optimization.

For my master's degree I studied traditional high-performance computing (HPC), mainly optimizing numerical models on heterogeneous CPU and GPU platforms, so I have a solid background in kernel-level optimization and system-level performance tuning. I did three AI Infra internships during the program and joined a large company through campus recruiting after graduating. I have worked on AI Infra there ever since.

I started this project to turn what I had picked up piecemeal at work into something organized, and to meet other people interested in AI Infra. I will keep writing notes and tutorials based on what I run into at work.