Features and benefits

The MLOps Workload Orchestrator solution provides the following features:

Pre-built ML workloads

The solution deploys with AWS CloudFormation templates to provision common ML workloads for Amazon SageMaker, such as model training, batch transform, real-time inference, and model monitoring.

Automatically deploy one or more ML workloads

Use the solution-provided API endpoint or Git commit mechanisms to automatically deploy ML workloads at scale.

Multi-account support

Easily deploy and promote ML workloads across different AWS accounts, such as development, staging, and production accounts. Multi-account support is for bring-your-own-model (BYOM) and model monitor pipelines.

Model monitoring

Monitor for deployed ML models and detect deviation in data quality, model quality, model bias, and model explainability.

Track Amazon SageMaker resources in a dashboard

Amazon SageMaker resources created by the solution, such as models, inference endpoints, models cards, and batch transform jobs, are automatically integrated with Amazon SageMaker Model Dashboard.

Notifications

Receive user notifications of pipeline outcomes through SMS or email.

Integration testing

You can run your own integration tests to ensure that the deployed model meets expectations.

Amazon SageMaker model registry

The solution provides an option to use Amazon SageMaker model registry to deploy versioned models. The model registry allows you to catalog models for production, manage model versions, associate metadata with models, and more.

Extend the solution with custom ML workloads

We designed the solution to be extensible and customizable. You can customize the pre-built CloudFormation templates or provide your own ML workloads templates.

Integration with Service Catalog AppRegistry and AWS Systems Manager Application Manager

This solution includes a Service Catalog AppRegistry resource to register the solution’s CloudFormation template and its underlying resources as an application in both Service Catalog AppRegistry and AWS Systems Manager Application Manager. With this integration, you can centrally manage the solution’s resources and enable application search, reporting, and management actions.