You are building a service that manages user data across product surfaces and uses machine learning to organize, retrieve, and prioritize the most relevant user information for downstream applications. The system must support both operational access patterns and ML driven decisions on top of evolving user data.
Design a scalable service for managing user data. What considerations would you make?
How you scope an ambiguous ML platform problemWhether you introduce retrieval and ranking instead of only storageHow you think about feature stores, serving, and driftHow you handle privacy, freshness, and fallback paths