Cloud data warehouses have fundamentally changed the way in which data analysts and scientists work, enabling them to quickly and efficiently analyze and model raw or transformed data. In this post, Ethan Rosenthal predicts that feature stores will have a similar impact on ML engineers – enabling them to quickly and efficiently build models using pre-existing features. He discusses some additional benefits of feature stores, including eliminating online-offline inconsistencies and providing frameworks to more easily transform raw data into high-quality features. However, he also predicts that feature stores could lead to a proliferation of models that intensifies the need for monitoring and governance.