legion

README

The Legion is an open source project dedicated to simplify automation of AI/ML products life cycle for small and large-scale enterprise systems.

Overview

• Modular Architecture
  • Separate components for different AI product life cycle phases: training, packaging and deployment
  • RESTful APIs and SDKs for all components to build extensions and plugins for external systems
  • Command line interface to interact with system components
  • Plugins for IDEs, Workflow Engines, CICD Engines
  • ML experiments tracking component
  • Logging, monitoring and alerting components
• Multicloud
  • Core system components deployment in K8S with Helm Charts
  • Deployment automation for major clouds: AWS, Azure, GCP with Terraform
• Security
  • Single sign-on (SSO) with OAuth2 or SAML2
  • System API endpoints secured with JWT tokens
  • Credentials manager for connectors to external systems
• Multiple AI/ML Toolchains
  • All major ML toolchains Scikit-learn, Keras, Tensorflow, PyTorch, H2O and others are supported
  • Integrated with MLFlow open source project

Documentation

Documentation sources

Changelog

Please find it here CHANGELOG

Backlog and Kanban Board

• Backlog
• Kanban Board

License

Apache 2.0

Contribution

If you are interested in contribution to project feel free to create an issue in backlog, create a pull request.