To scale up the operational analytics on assets data, AIH enables an automatic deployment of already trained models and harvesting value across different business areas.
Trained models that can be of data explorative, wrangling/processing, descriptive or predicting in nature can now be deployed in AIH via API. A trained models is configured in the form of a “job” when it is already containerized in Azure cloud.
Once the job is configured it can be executed in 3 different ways:
- scheduled (via dedicated end-point in AIH)
- on demand (via dedicated end-point in AIH
- event-based (via service bus subscription in AIH)
The output of the execution is also available via API.