AWS Certified Machine Learning – SpecialtyMachine Learning Implementation and OperationsEasy
A data science team is developing a fraud detection model. They need a centralized, version-controlled repository to store, share, and manage features used across multiple models and teams, ensuring consistency between training and inference. The solution must support both online (low-latency) and offline (batch) access patterns. Which AWS service is best suited for this requirement?
- AAmazon SageMaker Feature Store
- BAmazon S3
- CAWS Glue Data Catalog
- DAmazon DynamoDB
Show answer & explanationAnswer & explanation
Correct answer: A. Amazon SageMaker Feature Store
Amazon SageMaker Feature Store is a purpose-built service for storing, updating, and serving machine learning features for training and inference. It supports both online (low-latency) and offline (batch processing) access, ensuring consistency and reusability.
Why the other options are wrong
- B. Amazon S3 can store raw data and feature sets, but it lacks built-in capabilities for feature versioning, online serving, and consistent access patterns for ML.
- C. AWS Glue Data Catalog organizes metadata for data lakes but doesn't store the feature data itself or provide online serving capabilities for ML.
- D. Amazon DynamoDB is a key-value and document database that could serve features online, but it's not purpose-built for feature management, versioning, or offline access for ML workflows.
SageMaker Feature Store
A fully managed service in Amazon SageMaker that provides a centralized repository to store, retrieve, and share machine learning features for consistency between training and inference.
- Supports online (low-latency) and offline (batch) access.
- Ensures feature consistency for training and inference.
- Provides feature versioning and discoverability.
Memory trick: Feature Store keeps your ML data organized and ready for action.