mlc-ai/modern-gpu-programming-for-mlsys ? reverse-engineered prompt

Build me a clean online book site that teaches modern GPU programming for machine learning systems, starting from how the GPU works and building up to advanced kernels. It should read like a real tutorial, with chapters for the basics, memory layout, tensor cores, async data movement, GEMM, and flash attention, and it should include a reference section too.

I want the whole thing to be easy to read in the browser, with a simple navigation sidebar, code examples, and nice math and diagrams where needed. Please make it work as a static documentation site and set it up so it can be built locally and published online. If anything needs current docs or best practices, look them up online. Also keep in mind there should be an English site and a Chinese version.