注册一亩三分地论坛,查看更多干货!
您需要 登录 才可以下载或查看附件。没有帐号?注册账号 
x
Netflix Data Science and Engineering Intern (Experimental and Causal Inference)
一共两轮面试。
第一轮为自己的research介绍,大概15分钟,面试官会提出自己的疑问 (疑问不一定是你的的
研究重点氫,比较随机)。紧接着是 casestudy,本人被问到了两个。总共用了 45分钟就结束了,最后15分钟自己提问/聊天。
第二轮是本人没进,但邮件里表示是 15min
research presentation+15Q&A,再有 30min现场回答2个HM如何解决他们目前组里的问题。
哎还是不太会答 product问题,以及平时research 没有深入研究,一旦偏离自己的研究重点就不会...
以下是具体一轮面经内容
Simple self-introduction 3 min
Research
* Introduce your research
* How you estimate spill-over effect
* What are your research assumptions and why
* Adaptive: user crrect, conclude the 3 are significant.
2. Segment the 7, and see which are still significant to understand the user behavior.
2. How is the false negative change?
1. My answer: if 0.05 decrease to 0.01, you are more likely not to reject H0 and thus more likely to accept H1. Then if h1 is true, then power will increase. If H1 is not true, type II also increases. |