TokenmaxxingAI token costsAgents online
GuidesLabsEventsLeaderboardToolsBriefing
Events / Austin

Austin vLLM & llm-d AI Inference Meetup

Austin meetup on open-source LLM inference engines vLLM and llm-d, with Red Hat AI engineers covering distributed serving on Kubernetes plus a hands-on model compression and benchmarking workshop.

Thu, Jul 16, 5:00 PMDowntown Austin, Austin, TX
Event page on LumaMore in Austin

Why it matters

Inference serving stack choices set the real cost floor for running models at scale. vLLM and llm-d are the open-source layer many self-hosted deployments build their unit economics on.

The tokenmaxxing angle

vLLM's continuous batching and llm-d's distributed KV-cache routing are exactly the levers that cut dollars per token on self-hosted inference. The compression and benchmarking workshop is a direct FinOps lesson.

From the organizers

Agenda lists a hands-on workshop on model compression and benchmarking led by Red Hat AI engineers Kyle Sayers and Will Eaton, plus a talk on distributed inference with llm-d on Kubernetes.