Kortical SDK Documentation

The Kortical SDK primarily provides a way for users to drive the platform from code and to create and deploy custom ML apps using the command line interface. The primary uses for the Kortical package are to drive rapid model experimentation from code and to create, deploy and manage production machine learning apps and services. On top of this the package provides a number of very useful features for time series feature creation, data anonymization and more.

The package provides a few areas of functionality.

• Platform API - drive the Kortical platform fully through code (also includes advanced API, experiment framework and superhuman calibration).
• Reporting - send notifications (e.g training results, server errors) to Slack.
• Datasets - get started right away with toy datasets you can upload to the platform.
• Feature engineering - includes time series transforms.
• Anonymization - obfuscate sensitive data.

To get started with the SDK, refer to the "Setup" page for installation, or jump straight to the guide "ML Ops - Deployment". For full documentation, refer to the following sections.