SYSTEM Cited by 1 source

Amazon SageMaker Ground Truth

Amazon SageMaker Ground Truth is AWS's managed data-labeling service within the SageMaker AI family: defines labelling jobs + assigns them to a workforce (public / private / vendor) + returns structured annotations suitable for model training. Supports image bounding boxes, semantic segmentation, named entity recognition, and custom templates.

Stub — expand as dedicated sources arrive.

Seen in

• sources/2026-04-01-aws-automate-safety-monitoring-with-computer-vision-and-generative-ai — Labelling-job substrate for the AWS safety-monitoring solution. A Step Functions workflow triggered by EventBridge creates Ground Truth jobs on a configurable cadence, prioritised by the data-driven curation pipeline (Athena-surfaced false-positive hotspots + low-confidence inferences + Claude-detected class imbalance). Lambda post-processing transforms completed jobs into training-ready format; metadata in DynamoDB, annotations in S3.