AWS SageMaker
AWS's ML platform — data labelling, training jobs, hyperparameter tuning, model registry, deployment, monitoring in one stack.
Launched at re:Invent 2017. SageMaker Studio is the web IDE. SageMaker JumpStart has pre-built foundation models (including many from Bedrock). SageMaker HyperPod (2023) builds 1000+ GPU clusters for LLM training. Pipelines for end-to-end workflows.
Competitors: Vertex AI (GCP), Azure ML, Databricks, Snowflake. SageMaker is complex but unmatched in AWS integration. Sweet spot: enterprises that want "managed everything" inside an AWS account.