A data scientist defined transformations and feature engineering in SageMaker Data Wrangler and saved them to SageMaker Feature Store. New historical data lands periodically in S3 and must be transformed and added to the online feature store, prepared for training and inference using native integrations. Which solution requires the least development effort?

The correct answer is to use Amazon EventBridge to trigger a predefined SageMaker pipeline. EventBridge can directly detect new objects in S3 and trigger a SageMaker pipeline, which can execute the Data Wrangler flow for transformations and write to Feature Store. This approach leverages native AWS integrations, minimizing custom code and development effort. Using an AWS Lambda function to run a pipeline would require writing and maintaining Lambda code, increasing development overhead. Similarly, AWS Step Functions would involve defining and managing a state machine, which is more complex than a direct EventBridge trigger for this specific use case. Apache Airflow, while powerful for orchestration, introduces an external tool that needs to be set up, configured, and maintained, adding significant overhead compared to native AWS services.