SYSTEM Cited by 1 source

SageMaker Batch Transform Job

SageMaker Batch Transform Job is a managed compute primitive inside SageMaker AI that applies a trained model to a large batch of input records. Model artifact + input S3 location → inference container runs across all inputs → predictions written back to S3. No online endpoint is required; the infrastructure is spun up for the job and torn down on completion.

Stub page — expand as deeper internals are ingested.

Seen in

• sources/2025-06-29-zalando-building-a-dynamic-inventory-optimisation-system-a-deep-dive — canonical disclosure as the offline / batch delivery mechanism of Zalando ZEOS's replenishment recommender; runs the Monte Carlo + black-box optimiser daily across all merchants × all articles using the latest inventory settings from the offline feature store.
• — named as the batch-predictions stage of Zalando Payments' risk-scoring zFlow workflow (earlier in the Zalando canon than the 2025-06-29 post).

Related

• systems/aws-sagemaker-ai — parent product.
• systems/sagemaker-training-job · systems/sagemaker-processing-job — sibling primitives.
• companies/zalando