Article48 min read
Writing Kubernetes Controllers
Learn how Kubernetes controllers work, from watches and queues to reconciliation, scaling, testing, and production use.
Articles
Detailed but simple explanations of how production systems work, where they fail, and which tradeoffs matter.
Article12 min read
Prefill and Decode Need Different Resources
Prefill processes the prompt. Decode generates one token at a time. Learn when running them on separate workers improves speed and capacity—and when it does not.
Article14 min read
What Actually Happens When You Send a Prompt to an LLM?
Follow one LLM request from the API gateway to the GPU and back. Learn what prefill, decode, and the KV cache do along the way.