AIML accelerator recipes

The recipes in this accelerator are organized in the following folders:

• Functions
• Bootstrap
• AWS SageMaker
• AWS S3
• API Endpoint recipes

---

Functions

AIML | FUNC-005 | CSV Validator

This recipe defines the validation logic. The AIML SDK actions use this logic to validate the data against predefined configurations in the AIML | Data Fields lookup table.

---

Bootstrap

AIML| REC-009 | Load Model Schema

Use this recipe to automatically load your schema into the AIML Data Fields lookup table. Inspect the lookup table after using this recipe to ensure that the recipe loaded the correct datatype and validation rules.

AIML | FUNC-012 | Get Workato Data Type

This recipe works as a support recipe for AIML | REC-009 | Load Model Schema. It identifies the schema's data type.

---

AWS SageMaker

AIML | FUNC-003 | Get Sagemaker Prediction

Calls the Sagemaker InvokeEndpoint function to generate predictions. See Sagemaker documentation for more information.

AIML | FUNC-011 | Train ML Model Sagemaker

Calls Sagemaker CreateAutoMLJob endpoint to create a new AutoML training job. See the CreateAutoMLJob documentation for details.

---

AWS S3

AIML | FUNC-004 | Stage Training Data in S3

The recipe connects with S3 to get the source and target data.

---

Endpoint recipes

AIML | API-001 | Stage Training Data

Stage the training data from the source to a target location while validating it against the rules you define in the AIML | Data Fields lookup table.

AIML | API-002 | Get Prediction

Pass new, unseen feature data to the trained model to generate a prediction. The accelerator validates the data against the rules you define in the AIML | Data Fields lookup table and then passes it to the ML platform to return the prediction data.

AIML | API-002 | Train ML Model

This recipe trains a new ML model. This recipe takes the ML Model configuration, source, and target training data location from the AIML | Model Profile lookup table. It reads the data from the source, validates it, and then passes it to the ML Model to train a machine learning model.

AIML | API-008 | Get Model Identifiers and Schema

Reads the Data Fields and Model Identifier defined in the lookup tables and loads them in the actions.

---