Amazon SageMaker AI documentation hub — the center for all data, analytics, and AI in AWS.
Notes
| Topic | Description |
|---|---|
| SageMaker-Overview | What, why, benefits, pricing, constraints |
| SageMaker-Data-Tools | Unified Studio, notebooks, Data Agent, Lakehouse |
| SageMaker-Models | JumpStart, customization, HyperPod, inference |
| SageMaker-Governance | Catalog, lineage, quality, access control |
| SageMaker-Features | Studio, MLflow, MLOps, Feature Store, Clarify |
Quick Reference
What is SageMaker AI?
Amazon SageMaker AI is AWS’s fully managed ML platform providing:
- Complete ML lifecycle (prepare → train → deploy → monitor)
- 600+ foundation models via JumpStart
- Distributed training on thousands of GPUs (HyperPod)
- Enterprise governance via SageMaker Catalog
Key Components
Amazon SageMaker
├── SageMaker AI → ML model development
├── Unified Studio → Single IDE for all workloads
├── JumpStart → Model hub (600+ FMs)
├── Catalog → Data & AI governance
├── HyperPod → Distributed training
└── Lakehouse → Unified data storage
Access
🔴 AWS account required 🟡 Free tier available for Unified Studio features
SageMaker vs Bedrock
| Aspect | SageMaker AI | Bedrock |
|---|---|---|
| Focus | Custom model development | Pre-built foundation models |
| Skill level | Data scientists, ML engineers | All developers |
| Training | Full custom training | Fine-tuning only |
| Best for | Enterprise ML, custom models | Quick GenAI apps |