How would you design a scalable machine learning pipeline for a Boston Consulting Group client with high-frequency, low-latency data requirements?
