QuestionQ62

VerifiedDeployment and Orchestration of ML Workflows

A music-streaming company continuously sends song ratings from an application to an Amazon S3 bucket. The company wants to use the ratings as input for the training and inference of an Amazon SageMaker AI model.

The company has an AWS Glue Data Catalog configured with the S3 bucket as its source. An ML engineer must implement a solution that creates a repository for this data. The solution must ensure the data remains synchronized for batch training and real-time inference.

Which solution meets these requirements?

Explanation

Amazon SageMaker Feature Store provides a feature repository with an offline store for historical data used in model training and batch inference and an online store for low-latency feature retrieval during real-time model serving. When both stores are enabled for a feature group, they synchronize to avoid discrepancies between training and serving data. Feature Store supports batch ingestion from Amazon S3 sources and can use AWS Glue catalog metadata for its offline store.

Learn more

Community Discussion

No comments yet. Be the first to start the discussion!