Summary

In this chapter, we introduced you to the concept of deploying ML models outside of the cloud, primarily on an edge architecture. To lay the foundation for how to accomplish an edge deployment, we also examined what an edge architecture is, as well as the most important factors that need to be considered when designing an edge architecture, namely efficiency, performance, and reliability.

With these factors in mind, we explored how the AWS IoT Greengrass, as well as Amazon SageMaker services, can be used to build an optimal ML model package in the cloud, compiled to run efficiently on an edge device, and then deployed to the edge environment, in a reliable manner. In doing so, we also highlighted just how crucial the ability to manage and monitor both the edge devices, as well as the deployed ML models is to create an optimal edge architecture.

In the next chapter, we will continue along the lines of performance monitoring and optimization of deployed ML models.