最重要的是量化影响(用metrics和numbers):比如increasing revenue, customer acquisition, customer retention就是比较好的metrics,如果没有这些直接对business的影响,可以说improve team efficiency and productivity by building data pipelines,或者how a tool you built or methodology you developed was adopted by multiple teams。另外还要注意,不用谈到太多无关help business的细节。
原因是你大概率会被问到follow-up questions,提前准备好才不会大脑一片空白。比如Does it make sense to convert a continuous variable to a categorical variable? What if the distribution is long-tail? With a tree-based model, do you need to trim if it’s a random forest model?
5. 重视behavior表现。
.
面试官在你做presentation时也会评估the way you handle yourself,要显得自信,但也要避免显得arrogant 或者egotistical,一定要是个good listener,open to questions and suggestions。也就是说,承认别人提的问题很好,并给出解决方案。-baidu 1point3acres
很喜欢视频里的这句话,也分享给大家:A presentation in an interview is less about presenting a project and more about presenting yourself as a data scientist. . check 1point3acres for more.