Your question is Ensuring Distributed Data Integrity. 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 do you ensure data consistency and integrity across distributed database systems?
Explain how you would reason about consistency requirements, transactions, replication, conflict resolution, schema evolution, validation, recovery, and observability. Address how the design should protect downstream ML feature and training data from duplication, loss, corruption, stale reads, and training-serving skew.