AutoML Runner

The AutoML Runner automates the workflow of creating, training, and deploying machine learning models. It integrates seamlessly with the Data Science Lab (DSLab) module and allows you to import and operationalize models directly in a pipeline.

The AutoML Runner processes input data from one event, applies the trained AutoML model, and outputs enriched data (for example, with prediction columns) to another event.

Requirements

Before using the AutoML Runner, ensure that:

  • A model has been created and saved in the DSLab module.

  • Input and output events are available in the pipeline workflow.

  • Sufficient compute resources (CPU/GPU) are configured for model execution.

Using the AutoML Runner in a Pipeline Workflow

  1. Drag and drop the AutoML Runner component into the Workflow Editor.

  2. Create two events (input and output) and add them to the workspace.

    • Input data may originate from ingestion components, readers, DSLab scripts, or shared events.

  3. Connect the input event → AutoML Runner → output event.

  4. Select the AutoML Runner component to access its configuration tabs.

  5. Configure the fields in the Basic Information and Meta Information tabs.

  6. Save the component.

Once configured, the AutoML Runner consumes data from the input event, executes the trained model, and outputs the processed data with predictions to the output event.

Configuration

The AutoML Runner configuration is grouped into three sections:

  • Basic Information

  • Meta Information

  • Resource Configuration

Basic Information Tab

The Basic Information tab defines general properties and execution behavior.

Field
Description
Required

Invocation Type

Select execution mode: Batch or Real-Time.

Yes

Grace Period (sec)

Appears only for Batch mode. Time before the component shuts down gracefully.

Conditional

Deployment Type

Displays the deployment type of the component (pre-selected).

Yes

Container Image Version

Displays the Docker image version used (pre-selected).

Yes

Failover Event

Select an event to handle failover scenarios.

Optional

Batch Size

Maximum records processed per cycle (minimum 10).

Yes

Meta Information Tab

The Meta Information tab links the AutoML Runner to a project and model in the DS Lab module.

Field
Description
Required

Project Name

Name of the project containing the AutoML model.

Yes

Model Name

Name of the saved model within the project.

Yes

Saving the AutoML Runner

  • Click the Save Component (Storage icon).

  • A success message confirms that the component has been saved.

Example Workflow

  1. Ingest product sales data through a Reader component.

  2. Pass the data to the AutoML Runner, configured with a demand forecasting model.

  3. Write the enriched dataset (with predicted demand values) to an output Event.