Overcoming automation challenges with SageMaker Autopilot

In Chapter 1, Getting Started with Automated Machine Learning on AWS, we practically highlighted the challenges that ML practitioners face when creating production-ready ML models. By way of a recap, these challenges are grouped into two main categories:

• The challenges imposed by building the best ML model, such as sourcing and understanding the data and then building the best model for the use case
• The challenges imposed by the ML process itself, the fact that it is complicated, manual, iterative, and continuous

So, in order to better understand just how Autopilot overcomes these challenges, we must understand the anatomy of the Autopilot workflow and how it compares to the example ML process we discussed in Chapter 1, Getting Started with Automated Machine Learning on AWS.

Before we begin to use Autopilot, we need to understand that there are multiple ways to interface with the service. For example, we...