In this workshop you will explore the development cycle of machine learning model on AWS. In the first part, you will find a sample project fully developed in an ml.m4.4xlarge SageMaker notebook instance. On purpose, the notebooks are divided in different stages
- Exploratory analysis
- ETL to prepare training data
- Training the model with Hyperparameter Optimization
- Putting "new data" through a preprocessing pipeline to get it ready for prediction
- Batch predictions for new data
In the second part of this workshop we will implement this project in production automatizing it's execution using a combination of CloudWatch, Step Functions, Lambda, Glue and SageMaker.