Predictive Analytics for the Modern Enterprise

Chapter 8. Exploring AWS Cloud Provider Services for AI/ML

There are only a handful of large public cloud providers in the world. The ones you have probably come across are Amazon Web Services (AWS), Google Cloud Platform (GCP), Microsoft Azure, and Alibaba Cloud. While there is a lot of buzz around generative AI at the time of this writing, cloud providers have been helping enterprises accelerate their AI vision for a few years now. In this chapter, we will explore some AI/ML services provided by AWS and its parent company, Amazon.

To Cloud or Not to Cloud

Prior to delving into AI/ML cloud offerings, I’d like to discuss the value of dedicating time and resources to constructing machine learning (ML) solutions using public cloud provider services. Why would one choose to use AWS services (or any other cloud provider) instead of building their solution open source or in software that is cloud agnostic? Although this question may appear theoretical, it is frequently a significant consideration for chief technology officers and other technology executives. To understand this, we need to do a quick cost-benefit analysis, with “cost” referring not just to money, but also to time and control, among other things.

An AI/ML pipeline is a composite of multiple components that come together to serve the ML needs of an enterprise. For large companies, this pipeline needs to ingest, transform, and process large amounts of data from a multitude of sources. This requires significant investment ...