Databricks Generative AI Engineer Associate Question 218
Select 4You are tasked with deploying an endpoint for a basic Retrieval-Augmented Generation (RAG) application in Databricks. What are the correct steps needed to deploy this endpoint?
- A
Prepare the vector store by indexing the relevant documents.
- B
Train a custom language model from scratch for the application.
- C
Create a Databricks model serving endpoint for the RAG workflow.
- D
Develop the RAG pipeline, combining the retrieval and generation components.
- E
Test the endpoint by sending queries and validating the responses.
Show answer and explanation
Correct answers: A, C, D, E
Explanation
Deploying an endpoint for a basic RAG application involves preparing the vector store to support the retrieval component, setting up a model serving endpoint, and developing the RAG pipeline to integrate retrieval and generation. Testing the endpoint is also crucial to validate its functionality. Training a custom model from scratch is not required as pre-trained models are typically sufficient for a basic RAG application.
- A. Correct.
Correct. Preparing the vector store by indexing documents is a key step in enabling the retrieval component of a RAG application.
- B. Incorrect.
Incorrect. Training a custom language model from scratch is not necessary for a basic RAG application, as pre-trained models are typically used.
- C. Correct.
Correct. A model serving endpoint is required to deploy the RAG application for real-time use.
- D. Correct.
Correct. Developing the RAG pipeline is essential to integrate the retrieval and generation components effectively.
- E. Correct.
Correct. Testing the endpoint ensures that the application provides accurate and expected responses to user queries.