Deploy an LLM using llm-d

llm-d is a platform for inference serving that can be bigger than just using VLLM (in fact, it is used under the hood), but it can bring us some key features to scale up our AI platform. We need to verify the identify the pros and cons to install/maintain/usage.

To perform this spike, it's important to follow a GitOps approach, since ArgoCD will be our deployment platform on the HCMAI cluster.