Run Llama 3.1 locally with Ollama, benchmark it, then deploy a production API with vLLM and Docker.
This blueprint is queued. Steps will land here when ready.