节前面的，今天收到消息是挂了，贡献个面经给地里。\n面试官是国人小姐姐，职位在北美，整体体验感觉还可以，比较明显的感觉是面试官是looking for specific answer，要说中她心中的solution才算对。感觉BT的流程还是待改进，面试前跟recruiter三番五次确认面试内容，recruiter跟我斩钉截铁的说要面coding (python/R), 结果整个面试就考了道SQL题。\n\nProduct\n\nQ1： Assuming on Tiktok, there are the following 3 types of advertisers: each type of advertiser has different goals as listed. How do you design a system that picks among these 3 types of advertiser, which one to show to the user?\nGaming: App Install $10\nE-Commerce: Web Conversion $5\nBranding: Reach, Impression $3. 1point3acres\n\n\n问What’s the value of ads?\n答Different by advertiser type ($10, $5, $3)\n\nQ2： Let’s say we ran an e-commerce ads that’s targeting conversion rate, how do we prove the value of ads to the advertiser? (life in conversion rate)\n\nMetrics Question\nIf average time spent on Tiktok, and average time spent also increase for male on Tiktok, can we conclude that the overall time spent increases?\n. Waral dи,\nSQL\n以下内容需要积分高于 188 您已经可以浏览\n\ntable ad_id | user_id | timestamp\nFind the time difference between the last impression timestamp and the second last impression timestamp for the same user seeing the same ad\n\n\n又是一年面试季了，有帮助的话求个大米~. ----\n