https://iamartyaa.github.io/writings/ Long, illustrated notes on how things actually work. Every piece is built as its own page, in its own design system. en Tue, 01 Sep 2026 20:54:22 GMT https://iamartyaa.github.io/writings/serving-from-scratch/ https://iamartyaa.github.io/writings/serving-from-scratch/ Tue, 01 Sep 2026 09:00:00 GMT How a model becomes a service, one component at a time — one model, batching, streaming, many models — then a drawboard where you wire your own server and run it. LLM servingbatchingstreaming https://iamartyaa.github.io/writings/the-gemm-scrapbook/ https://iamartyaa.github.io/writings/the-gemm-scrapbook/ Sat, 29 Aug 2026 09:00:00 GMT One matrix multiply, seven CUDA kernels, and a 70× speedup — documented like a road trip, with the wrong turns left in. CUDAGPUperformance