You May Also Enjoy
Kimi K3: From KV Cache to Selective Memory
August 04 2026
I read through this Waterloo Intern thread on Kimi K3 and turned it into my own learning notes.
DSpark and DFlash: 10 Notes on Speculative Decoding
July 17 2026
I read Dmytro Dzhulgakov’s DSpark thread and the related papers, and here is my learning note.
pi-dynamo-provider: Wiring Pi Agents into Dynamo’s Observability Stack
May 21 2026
pi-dynamo-provider is a small (~650 LOC TypeScript) Pi extension that registers a dynamo model provider — letting Pi’s agentic CLI run against an NVIDIA Dyna...
Priority Scheduling on Dynamo + vLLM: Three Things That Have to Be True
May 15 2026
Short reference for “the next time I touch Dynamo + vLLM priority.” Long version lives in the project README.md; raw numbers are in results/ttft_by_priority....
