Integrate Label Studio with AWS Sagemaker

Overview

Label Studio integrates with Amazon SageMaker, a fully-managed machine learning service, to better serve the MLOps Community. Leverage Amazon SageMaker’s tools through Label Studio’s ML extension interface. As annotators work through datasets, functions can be triggered to AWS Lambda and SageMaker.

These functions can include actions to:

• Monitor model performance against ground truth annotations
• Notify experts when a new project is ready to be annotated
• Train models in an active learning pipeline
• Version datasets based on labeling activities

Webhooks in Label Studio are an excellent way to simplify and automate a SageMaker machine learning pipeline.

How Label Studio Connects with Amazon SageMaker

Label Studio connects to Amazon SageMaker through shared S3 storage and webhooks. You can import raw data from S3 into Label Studio for annotation, then use webhooks to automatically trigger a SageMaker training or retraining job whenever new labels are created. This setup keeps your models updated as new annotations arrive, and all data stays in your AWS environment. For a step-by-step setup guide, see Automate your ML pipeline with Label Studio and Amazon SageMaker.