Seldon Server is an open-source machine learning platform that enables deploying and managing ML models and recommendation engines at scale on Kubernetes.
Machine Learning Platform and Recommendation Engine built on Kubernetes
This tool is designed for data science teams and ML engineers who need to deploy, serve, and manage machine learning and deep learning models in production environments, either on-premise or in the cloud. It supports a wide variety of ML frameworks and provides APIs for prediction and recommendation, making it suitable for enterprises and startups aiming to operationalize ML workflows securely and efficiently.
This project is archived and no longer actively maintained; users are encouraged to migrate to Seldon Core for ongoing support and advanced features. Seldon Server requires a Kubernetes environment and familiarity with containerized deployments. Secure integration is facilitated via OAuth 2.0 and gRPC APIs. The platform supports complex algorithm orchestration without downtime, suitable for production-grade ML deployments.
Install a Kubernetes cluster (on-premise or cloud)
Follow the Seldon install guide at http://docs.seldon.io/install.html
Use Kubernetes manifests or Helm charts to deploy Seldon Server components
Configure your ML models and services according to the documentation
Access the CLI tool for managing deploymentsseldon-cli
Command Line Interface for configuring and managing Seldon Server deployments
API Predict endpoint
Send prediction requests to deployed supervised machine learning models
API Recommend endpoint
Use the recommendation engine APIs for user activity and content-based recommendations