179 questions
No questions match those filters.
← All questions
RAGMedium· Caching· Latency· System Design
A teammate wants every answer to fetch its sources live...
This is one of the questions in the full AI/ML interview bank. Pro unlocks all 1789 questions; Premium includes the same bank plus the highest daily Practice limit.
See plansAlways-live fetching is inefficient because it forces the system to pay the cost of a crawl on every single query. This is a significant waste of resources compared to traditional search engines that amortize this cost over many users. You should challenge the requirement by identifying the specific staleness window the product can tolerate, allowing you to cache results for sources that do not require real-time updates.