Build a ML Workflow for Scones Unlimited on Amazon SageMaker

Project Steps Overview

• Step 1: Data staging
• Step 2: Model training and deployment
• Step 3: Lambdas and step function workflow
• Step 4: Testing and evaluation
• Step 5: Optional challenge
• Step 6: Cleanup cloud resources

Deliverables

Train and Deploy a Machine Learning Model

• Set up a SageMaker Studio Environment
• Perform ETL (Extract, Transform, and Load)
• Train a ML model
• Construct an API endpoint associated with a model trained in Sagemaker

Build a full machine learning workflow

Author three Lambda functions

• 1st lambda is responsible for return an object to step function as image_data in an event
• 2nd lambda is responsible for image classification
• 3rd lambada is responsible for filtering low-confidence inferences

Author Step Function

• Compose Lambdas together in a Step Function.
• Export JSON that defines the Step Function
• Screenshot of the working Step Function.