Question Bank
RecSys interview questions
Recommendations must find relevant items while respecting product constraints. Review candidate retrieval, ranking and evaluation. Explain how cold starts, user feedback and the difference between offline metrics and online experiments affect your choices.
Showing 19 of 139 questions
How to use user interaction signals with search results?
How to integrate a new model into the existing system?
Do we know which metrics correlate with the company's revenue?
What should we do if the number of candidate results reaches a million?
What could be a drawback of a strategy that considers only additions to the cart?
What fallback would you provide if the model returned no recommendations?
What should we do to avoid recommending only similar products?
Are cart recommendations more of an item-to-item or an item-to-user task?
What would be the nearest neighbors of that ring with the luxurious gemstone?
What complementary products would I want to see alongside this ring?
Could we formalize this constraint and incorporate it into our A&N model?
What objective do we want cart recommendations to achieve?
How good is it if the recommendations contain only a ring?
Prepare for your next interview with Vibe Interview.
Download Vibe Interview