SYSTEM Cited by 1 source

Amazon SageMaker Batch Transform

Amazon SageMaker Batch Transform is AWS's managed batch- inference runtime within the SageMaker AI family. Runs a model over a large input dataset in S3, writing predictions back to S3 — no persistent endpoint, no per-request latency budget, scales on input size.

Stub — expand as dedicated sources arrive.

Seen in

• sources/2026-04-01-aws-automate-safety-monitoring-with-computer-vision-and-generative-ai — canonical substrate for large-scale synthetic-data generation. GLIGEN deployed as Batch Transform jobs producing a 75,000-image PPE dataset + 75,000-image Housekeeping dataset at 512×512 resolution with ground-truth bounding boxes embedded by construction. Parallel Python workers downstream of the Batch Transform output convert raw images + embedded annotations into YOLO training format and upload to S3 as training-ready datasets.