Summary
In this chapter, we discussed the terminology used in the feature store world, specifically terminology that relate to Feast. However, keep in mind that many of the existing feature stores use similar terminology, so if you are familiar with one, it is easy to understand the others. We also discussed how the point-in-time join works in Feast, along with the Feast fundamentals such as installation, initialization, project structure, and API usage. Finally, we explored the components of Feast and how the operationalization of a model works with Feast.
In the next chapter, we'll use Feast in the model we built in Chapter 1, An Overview of the Machine Learning Life Cycle, learn how it changes the way data scientists and engineers work, and see how it opens the door to new opportunities in feature sharing, monitoring, and easy productionization of our ML models.