Your question is Feature Store for Batch and Real-Time. Take a moment with it on the right.
Talk me through your thinking if you like. When you're confident, submit your answer and I'll grade it like a real screen (7/10 or better passes).
How would you architect a feature store to support both batch training and real-time inference?
Discuss the offline and online storage layers, feature computation and materialization, point-in-time correctness, consistency between training and serving, freshness guarantees, and operational monitoring. Explain the design choices, tradeoffs, and failure handling you would consider.