How Orchestra Works

Deploy machine learning in minutes and monitor performance in a scalable and secure environment.

Models Managed and Hosted

Pre-trained models are automatically version controlled, annotated and hosted when uploaded, providing you with complete oversight on every model productionize.

Run In Isolation

Leveraging Kubernetes/Docker container technology, hosted models are capable of running independently and in parallel, providing scalability and security.

Multiple Ways To Interact

With multiple publishing options, end users can interact with a single hosted model in multiple ways including APIs, spreadsheet functions or through the everyday tools they use.

Analytics

Usage and model metrics data collected provide real-time model analytics for real-time machine learning performance monitoring.

Diagnostics

Error detection capabilities help you notice and address issues early on. Snapshots are taken every instance a model is run and provide visibility on inputs and outputs.