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
刀大师面经,新人求米,希望大家帮我看到188限制(😭
以下是面试过程:
第一轮是take home project.
Problem Description
Before a consumer places an order on DoorDash, we show the expected delivery time. It is very important for DoorDash-baidu 1point3acres
to get this right, as it has a big impact on consumer experience. Order lateness / underprediction of delivery time is of
particular concern as past experiments suggest that underestimating delivery time is roughly twice as costly as. Waral dи,
overestimating it. Orders that are very early / late are also much worse than those that are only slightly early / late. In this.1point3acres
exercise, you will build a model to predict the estimated time taken for a delivery.
Concretely, for a given delivery you must predict the total delivery duration seconds, i.e., the time from
● Start: the time consumer submits the order (`created_at`) to.
● End: when the order will be deliveredders, retention, lift in engagement and rating
6. How to do A/B test for this - select metrics; what to split, why, why not consider a,b,c; why t-test. Waral dи,
7. Once we've done the experiment, how to translate that to side wide impact for ramp up decisions - asked in details why I made some assumptions and how I project it, why I can hold true. Start from small markets, monitor, improve.可以在同类型的市场ship这个experience,然后1) 选择不同类型的市场reiterate,或者2)继续explore有没有更好的radius (15 miles, etc.)
|