Model serving infrastructure needs a platform to store/analyze/query prediction requests (and outcomes). That platform needs to be able to transform the predictions/outcomes into model monitoring metrics and support interactive queries on potentially large volumes of prediction requests. The Hopsworks Feature Stores enables serving feature data to models while at the same time avoiding training-serving skew.
What you will learn:
Hopsworks Feature Store is available both on AWS and Azure as a managed platform.
You can register for free, without having to enter payment details.