amazon-sagemaker-from-idea-to-production  by aws-samples

Jupyter notebooks for SageMaker MLOps, from idea to production

created 2 years ago
400 stars

Top 73.4% on sourcepulse

GitHubView on GitHub
Project Summary

This repository provides a structured, six-step guide for moving machine learning projects from initial idea to production using Amazon SageMaker. It targets ML engineers and data scientists seeking to implement MLOps best practices, offering a practical, hands-on approach to operationalizing ML workflows.

How It Works

The project guides users through a series of Jupyter notebooks, each building upon the last. It starts with local experimentation and progressively integrates SageMaker's MLOps capabilities, including SageMaker Pipelines for workflow orchestration, Feature Store for managing ML features, Model Registry for versioning, MLflow for experiment tracking, and Model Monitor for performance oversight. This incremental approach allows users to gradually adopt advanced MLOps features.

Quick Start & Requirements

  • Install/Run: Clone the repository (git clone https://github.com/aws-samples/amazon-sagemaker-from-idea-to-production.git) and execute the 00-start-here.ipynb notebook within a SageMaker Studio environment.
  • Prerequisites: An AWS account is mandatory. A SageMaker domain and user profile are required, which can be provisioned via the AWS console or a provided CloudFormation template (sagemaker-domain.yaml). Specific IAM policies (AmazonSageMakerFullAccess, AWSCloudFormationFullAccess, AWSCodePipeline_FullAccess, AmazonSageMakerPipelinesIntegrations, AWSCodeStarFullAccess) must be attached to the execution role. A public VPC is assumed for the CloudFormation template.
  • Setup Time: Approximately 15 minutes for CloudFormation stack creation, plus time for SageMaker domain setup and notebook execution.
  • Resources: Links to AWS workshop documentation and SageMaker Studio setup guides are provided.

Highlighted Details

  • Demonstrates end-to-end MLOps lifecycle with SageMaker features.
  • Includes optional assignment notebooks for deeper hands-on practice.
  • Covers advanced topics like CI/CD integration, A/B testing, and model monitoring.
  • Uses the UCI Direct Marketing dataset for practical examples.

Maintenance & Community

This repository is part of the aws-samples organization, indicating official AWS backing. Further community engagement or roadmap details are not explicitly mentioned in the README.

Licensing & Compatibility

  • License: MIT-0.
  • Compatibility: Permissive license suitable for commercial use and integration with closed-source projects.

Limitations & Caveats

The CloudFormation template for domain creation is limited to us-east-1, us-west-2, and eu-central-1 regions. Users must ensure their AWS account has a public VPC configured if using the template.

Health Check
Last commit

1 week ago

Responsiveness

1 week

Pull Requests (30d)
0
Issues (30d)
0
Star History
59 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.