Skip to content

1. Learning Objectives

GENAIOPS: Ideation is part of the Customization stage. (click to view figure)

GenAIOps toolchain

In the previous sections we completed model selection and model exploration:

  • We learned to use the model catalog and benchmarks to "select" a model
  • We learned to use model deployment and chat playground to "explore" the model

We found the model works for our needs but we do need to ground the responses in our data! It's time to focus on model customization with an emphasis on Retrieval Augmented Generation (RAG) where we add our data to ground model responses. Let's do that, next!

By the end of this section you should know how to:

  • Create and connect an Azure AI Search service to an AI project
  • Create or update RBAC (role access permissions) for resources
  • Add your data to the chat model playground for grounding
  • Test your grounded chat model interactively in playground
  • Deploy that chat prototype as a standalone web application
  • Use the deployed application (with history) or update it