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
本帖最后由 匿名 于 2020-1-15 05:29 编辑 .1point3acres
网申之后一个月内, HR 会发邮件询问兴趣方向,有三个
1. program evaluation/experiment. .и
2. empirical IO
3. time series/forecast
选了之后 HR 会安排技术电面。
我选了3,第一轮电面 30分钟
一道 behavioral。
之后 technical:
1. 他们有不同 product 的 sales data time series,如果你要做 forecast,讲讲怎么做,具体步骤是怎样的。中间会提出很多技术问题,一步步交流怎么解决。
2. 上题数据一样,怎样 estimate prime day effect?
第二轮是 codepad 面 30分钟
再一道 behavioral
之后是 coding,跟第一轮 technical 很像。只是具体给了数据,然后考你用语言(不限语言)会不会把分析和预测做出来。没有太多 guideline,自由发挥,但要快,时间很紧。
The first dataset, “Enrollment_Info,” is at the county-insurer level where each county is identified by a unique “FIPS_Code” and each insurer is identified by a unique “Contract_ID”. This dataset contains the number of enrollees for each insurer within each county.
. From 1point 3acres bbs---------------------------------------------------------------------------------------------------
County | State | Contract_ID | Organization_Name | FIPS_Code | num_Enrolled
---------------------------------------------------------------------------------------------------
Autauga | AL | S0522 | SYMPHONIX HEALTH INSURANCE, INC. | 1001 | 50
Autauga | AL | S1030 | BCBS OF ALABAMA & UTIC INSURANCE COMPANY | 1001 | where possible.
4. Time Series (TS):
Assume that we give you a dataset that contains the number of people eligible to be enrolled and actually enrolled (across all insurers) within each county, for each month over the last two years. Construct a forecast of each state’s penetration over the next 3 months.
i. What type of model would you run? Implement it and explain your reasoning.
ii. Run any additional checks that are necessary to implement this forecast model.
iii. Discuss assumptions you must make in order to run the analysis and check them where possible.
iv. What additional data would you ideally want to include in this analysis?
第二轮 coding 资料这么详细具体,求交流,求大家的思路和做法,求题解。
应该没有人在地里发过这职位的面经。
求高评分和积分!
|