Databricks Generative AI Engineer Associate Question 177
Select 4You are tasked with deploying a generative AI application on Databricks that uses a pre-trained large language model (LLM) to generate customer support responses. The application requires real-time inferencing with low latency. Which of the following steps should you take to assemble and deploy the application successfully?
- A
Deploy the model as a REST API endpoint using Databricks Model Serving.
- B
Use Databricks Jobs to schedule batch inferencing tasks for the application.
- C
Set up feature engineering pipelines for real-time data processing.
- D
Leverage MLflow to track the model version and its deployment status.
- E
Use a custom cluster with GPU instances for optimal performance.
Show answer and explanation
Correct answers: A, C, D, E
Explanation
To assemble and deploy a generative AI application for real-time inferencing, you must ensure low latency by deploying the model as a REST API endpoint and using GPU instances for high-performance inferencing. Real-time data processing pipelines are crucial to preprocess input data, and MLflow must be leveraged to manage the model lifecycle and deployment status. Batch inferencing is not applicable in this scenario, as the application requires immediate responses.
- A. Correct.
Correct: Deploying the model as a REST API using Databricks Model Serving is essential for real-time inferencing, which requires low latency.
- B. Incorrect.
Incorrect: Batch inferencing is not suitable for real-time inferencing applications as it introduces delays.
- C. Correct.
Correct: Feature engineering pipelines are critical for processing input data in real time before passing it to the model.
- D. Correct.
Correct: MLflow is a key tool for managing model versions and deployments in Databricks, ensuring consistent and trackable deployment workflows.
- E. Correct.
Correct: Using GPU instances for the cluster ensures optimal performance for inferencing tasks with computationally intensive LLMs.