Summary
In this chapter, we set out with the aim of trying out a use case, namely telecom customer churn prediction using a dataset available from Kaggle. For this use case, we used a managed SageMaker Feature Store, which was introduced in the last chapter. In the exercise, we went through the different stages of ML, such as data processing, feature engineering, model training, and model prediction. We also looked at a feature monitoring and model monitoring example. The aim of this chapter was not model building but to showcase how to use a managed feature store for model building and the opportunities it opens for monitoring. To learn more about feature stores, the apply conference (https://www.applyconf.com/) and feature store forum (https://www.featurestore.org/) are good resources. To stay updated with new developments in ML and how other firms are solving similar problems, there are a few interesting podcasts, such as TWIML AI (https://twimlai.com/) and Data Skeptic (https...