How AWS makes automating the ML development and deployment process easier
The focus of the remaining chapters in this book will be to practically showcase, using hands-on examples, how the ML process can be automated on AWS. By expanding on the Age Calculator example, you will see how various AWS capabilities and services can be used to do this. For example, the next two chapters of this book will focus on how to use some of the native capabilities of the AWS AI/ML stack, such as the following:
- Using SageMaker Autopilot to automatically create, manage, and deploy an optimized abalone prediction model using both codeless as well as coded methods.
- Using the AutoGluon libraries to determine the best deep learning algorithm to use for the abalone model, as well as an example for more complicated ML use cases, such as computer vision.
Parts two, three, and four of this book will focus on leveraging other AWS services that are not necessarily part of the AI/ML stack, such as the following:
- AWS CodeCommit and CodePipeline, which will deliver the abalone use case using a Continuous Integration and Continuous Delivery (CI/CD) pipeline.
- AWS Step Functions and the Data Science Python SDK, to create a codified pipeline to produce the abalone model.
- Amazon Managed Workflows for Apache Airflow (MWAA), to automate and manage the ML process.
Finally, part five of this book will expand on some of the central topics that were covered in parts two and three to provide you with a hands-on example of how a cross-functional, agile team can implement the end-to-end Abalone Calculator example as part of a Machine Learning Software Development Life Cycle (MLSDLC).
