A feature platform is a feature store that also provides support for a domain-specific language (DSL) to define feature logic and feature pipelines. Feature platforms include an orchestration engine for both feature pipelines and on-demand features.
Feature platforms can provide benefits of a feature store such as centralization, standardization, and scalability. The DSL lowers the barrier for defining and productionizing common feature logic/pipelines. However, the DSL for feature logic/pipelines limits the scope of available feature transformations, data sources, feature engineering frameworks, and orchestration support. In addition to a feature platform, you will need platforms for training pipelines and inference pipelines. In practice, this means you may have different orchestration engines for different ML pipelines (e.g., the feature platform for the feature pipeline, and Airflow for training and batch inference pipelines).