Seldon Deploy is made up of several interacting components that the application uses. The core design is illustrated in the diagram below.
Seldon Deploy manages the running of our open source core components Seldon Core, KFServing, and Seldon Alibi. It also manages a range of over components to provide a full machine learning deployment platform.
The following sections will discuss the core parts of the architecture.
Serving technologies we support
User Authentication
Machine Learning Model Explanation
Source Control as the source of truth
End to End Machine Learning on Kubernetes
Real time analytics for ML models
Log prediction requests and responses for analysis