Had a 45 minute techical phone interview [got extended to 60 minutes roughly]
Problem-1 (Design Problem)
Assumig peta-bytes of documents data stored in a local, make <key, value> of all documents that can be indexed efficiently (something like Google Search) using reverse indexing method (https://www.geeksforgeeks.org/inverted-index/) Interviewer was providing helpful inputs to validate if I was going right.
Extension
Ater that asked me, what could be corner cases (memory complexity and runtime complexity) from a programmer perspective and given that data would be increasing at a huge scale down the line -- basically how to make the approach scalable.