Honestly, way too many people have been copying interview notes over to Xiaohongshu lately, so I've set some limits on mine — I don't want to be scrolling xhs and suddenly run into my own interview notes, lol.
AR: Felt like a new question — hadn't seen it in recent interview notes. The question was: suppose you're on the ads team and you built a new ad recommendation model, how would you go about figuring out whether it's a good idea (launch or not / experimentation). At the end they drew a graph and asked how I'd improve it — the y-axis was ad CTR, with two lines drawn for control and treatment.
AE: A DAU question. First they asked what the distribution of comments per DAU looks like. Then: if you resample 10k users and repeat that 100k times, what would the distribution of the group means look like? If we increase the 10k to 20k, how would that distribution change? We also went through what the mean, mode, average, and p95 would look like — what the values would be, and whether they'd be float or integer.
SQL: The shop visibility question.
BQ: Don't really remember it — I thought it was pretty easy. I've already been rejected.
Lately I've read way too much news about Meta's layoffs and the rat race there, and honestly I've kind of given up on it. So I went into the interview pretty relaxed — if I don't get it, so be it. Mental and physical health matter more than anything.
Edit: One more thing — after the AE round they also asked a case question, mainly root cause analysis: a KPI suddenly dropped, how would you go about finding where the problem is, and how would you validate your hypothesis.
Discussion
Loading comments…