Effortlessly orchestrate your ML pipelines on AWS with ZenML

Leverage the power of AWS for your machine learning workflows by integrating it with ZenML. This integration allows you to seamlessly run your ZenML pipelines on AWS infrastructure, taking advantage of its scalability, reliability, and extensive set of ML services. Whether you're looking to train models, deploy them, or manage complex workflows, combining AWS with ZenML streamlines your MLOps processes.

Features with ZenML

Seamlessly orchestrate ZenML pipelines on AWS SageMaker for efficient training and deployment
Utilize AWS S3 for convenient storage and access to datasets and artifacts
Leverage AWS EC2 instances for scalable compute resources in your ML workflows (via AWS Skypilot integration)
Integrate with various AWS services, such as ECR for container registry
Simplify MLOps processes by combining the power of AWS with ZenML's pipeline management capabilities

Available integrations

‍

Main Features

Comprehensive suite of machine learning services, including SageMaker for model training and deployment
Highly scalable and reliable cloud infrastructure for running ML workloads
Secure and compliant environment for handling sensitive data and models
Extensive set of tools for monitoring, logging, and debugging ML workflows
Wide range of compute instances and accelerators to optimize performance and cost

‍

How to use ZenML with

AWS

# Register a ZenML stack using existing cloud components in AWS Cloud.
zenml stack register <STACK_NAME> -p aws

# Deploy needed ZenML stack components into your's AWS Cloud.
zenml stack deploy -p aws -n <STACK_NAME>

The code example demonstrates how to set up and use the AWS stacks with ZenML. It offers two options: register a stack using existing cloud resource or deploy a stack from scratch. User can choose the most valid one based on personal situation.

Additional Resources

Set up a minimal AWS stack

AWS Service Connector Documentation