As customers move from building data lakes and analytics on AWS to building machine learning solutions, one of their biggest challenges is getting visibility into their data for feature engineering and data format conversions for using AWS SageMaker. In this workshop, we demonstrate best practices and build data pipelines for training data using Amazon Kinesis Data Firehose, AWS Glue, and Amazon SageMaker, and then we use Amazon SageMaker for inference.