A music streaming company streams song ratings into an S3 bucket and has an AWS Glue Data Catalog pointing to that bucket. The company wants a repository for these ratings that stays synchronized for both batch training and real-time inference. Which solution satisfies this requirement?

SageMaker Feature Store is designed for exactly this scenario: a centralized, managed repository for machine learning features that supports both batch and real-time access. Ingesting data from S3 into Feature Store allows for consistent feature definitions, versioning, and low-latency retrieval for inference, while also being available for batch training. Applying tags and indexes further enhances discoverability and query performance. Amazon Athena is a query service for data in S3, not a feature store for ML. While it can group data, it doesn't provide the real-time serving capabilities or feature management benefits of SageMaker Feature Store. AWS Lake Formation is a security and governance layer for data lakes, not a feature store itself. It manages permissions but doesn't offer the operational features required for ML model serving. SageMaker Data Wrangler's Generate Data Insights is for exploratory data analysis and feature engineering, not for creating a persistent, synchronized feature repository.