Managed Machine Learning Services (Tier 2)
The ML services tier provides a more customized approach for customers who want more control over their ML solutions without having to manage underlying infrastructure. Amazon SageMaker is the key offering in this tier, providing a comprehensive platform for the entire machine learning workflow.
Amazon SageMaker​
Amazon SageMaker is a fully managed service where you can build, train, and deploy your own ML models without worrying about infrastructure. Its integrated development environment (IDE) provides simplified access control and transparency over your ML projects, enabling tracking, visualization, debugging, and monitoring of workflows all within one environment.
| Service Name | Logo | Key Attributes | Primary Use Cases |
|---|---|---|---|
| Amazon SageMaker |  | - Fully managed ML platform. - Integrated development environment (IDE). - Offers access to hundreds of pre-trained models. | - Building, training, and deploying custom ML models. - Tracking ML experiments and visualizing data. - Debugging and monitoring ML workflows. |
Key Benefits of SageMaker​
- Choice of ML Tools: Increase innovation with different tool choices; data scientists can use the IDE, and business analysts can use the no-code interface.
- Fully Managed Infrastructure: Focus on ML model development while SageMaker provides high-performance, cost-effective infrastructure.
- Repeatable ML Workflows: Automate and standardize MLOps practices and governance across your enterprise to support transparency and auditability.
ML Frameworks and Infrastructure (Tier 3 Context)​
For organizations with highly specialized needs and in-house expertise, AWS also supports the use of ML frameworks and underlying infrastructure (Tier 3). This allows for complete control over the ML training process.
- ML Frameworks: Software libraries like PyTorch, Apache MXNet, and TensorFlow provide experienced ML practitioners with pre-built, optimized components for building models.
- AWS ML Infrastructure: Services like ML-optimized Amazon EC2 instances, Amazon EMR, and Amazon Elastic Container Service (Amazon ECS) support these custom solutions, offering high performance and flexibility for advanced ML workloads.