Welcome!

I’m Scott, an undergrad in Zhejiang University’s Turing Class, currently doing research at ZIPLAB under Prof. Zhuang Bohan.

I work on efficient AI, but what really pulls me is the machinery underneath it: AI infrastructure and HPC.

Lately I’ve been researching KV cache design for video generation models.

Selected Writing

• Lab1 - HPC summer lab on the numerical stack, Slurm/HPL, shared storage, K3s, and observability.
• WGMMA 与异步执行范式 - a seed note on Hopper-style async GPU kernels.
• Ozaki Scheme - using low-precision Tensor Cores to emulate high-precision GEMM.
• Hello Quartz - a short first post for the new garden.

A few places you might wander:

• Blog - longer writing on systems, learning, and whatever I’m chewing on.
• Courses - course-bound notes, reports, and MOCs.
• Notes - a garden of half-formed thoughts, mid-read papers, and things I’m still figuring out.
• GitHub - code and project traces.
• Email - the simplest way to reach me.

— Scott

---