Consolidating Austin's Elite into a monorepo and rehearsing an isolated Coolify deployment down to the proxy rules.
A late night stabilizing a fleet of model profiles, capping inference hangs, and promoting a faster quantized fleet.
Dockerfiles, compose files, nginx config, and entrypoints—standardized so every surface deploys the same way.
Building a UE5 dedicated server image, fighting wine prefixes, and writing down the debugging war stories.
One session took ArcLLM from an empty repo to streaming chat, tool calls, the responses API, persistence, and worker queues.