[MIN-103] Deploy OpenLLM on AKS

[dudeperf3ct]

In this PR, we deploy flan-t5-xl on AKS. For this, we will provision two VMs. One VM on AKS acts as CPU node (or a system node) and a GPU node (which is a VM with T4 GPU instance).

The task can be broken down into following todo list

• Update the deployment in infrastructure folder to use openllm as an alternative to seldon deployment
• Create a Dockerfile for running the server
• Update the application logic to use openllm
• Update the Readme of MindGPT with instructions to deploy using openllm. For this we will create a separate section that outlines shortcoming of seldon approach for xl model and steps required for deploying xl model using openllm.