
Amazon SageMaker Data Wrangler introduces S3 manifest file support, inference artifact generation, and JSON format integration to streamline data preparation and model training workflows.


<div><p>This article announces new features in Amazon SageMaker Data Wrangler that enhance data preparation and ML model training workflows.</p><ul><li>S3 manifest file support enables training on entire large datasets, not just subsets</li><li>Inference artifacts can now be generated programmatically from processing jobs</li><li>JSON format support for both batch and real-time inference endpoints</li><li>Seamless integration with SageMaker Autopilot for automated model training</li><li>Simplified data transformations (one-hot encoding, PCA, imputation) for production inference</li><li>Code-first and no-code paths available for MLOps workflows</li></ul><p>These enhancements streamline data preparation, reduce manual effort, and enable efficient end-to-end ML workflows with SageMaker Data Wrangler.</p></div>


Related articles