Your question is Feature Engineering on Big Data. 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).
You're working on an ML pipeline and need to prepare features from data that is too large for a single machine. You want an approach that supports both training and repeated production use.
What techniques would you use for feature engineering on big data?