Your question is A/B Testing for Recommendation Algorithm. 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).
Explain how you would set up an A/B testing framework to evaluate a new recommendation algorithm, including sample size calculation and handling network effects.
Discuss the hypothesis, primary metric, guardrails, minimum detectable effect, randomization unit, and duration. Show the sample-size calculation with explicit assumptions, describe the analysis and peeking policy, and explain how network interference or violations of SUTVA could change the design. State a clear ship, iterate, or do-not-ship rule.