Interview Linkedin(电话面试), FB(电话面试), Uber(电话面试)

有些是几个月前面试的,大概记得some

  1. linkedin:
    Machine learning part:
    Q1. Overfitting , Q2. Regularization ,Q3. Naïve bayes classifier ,Q4. Talk about your research and your methods used
    Coding: Give a m face dices, and the probability of each face. Generate a N-length sequence of the faces, so that overall they follow the probability distribution. (use accumulated probability distribution + binary search)
  2. Uber:Case questions: 1) how to predict that a driver will cancel a ride, what kind of features you need 2) how to predict that a rider will cancel a ride, what featuresCoding question: Leetcode questions, Given [0, 1,1,1,0,1], return [0,0,1,1,1,1]
  3. FB;Coding question: Given list A= [0, 0, 0,1, 1, 1,0,0,1,1], here 0 means national holidays you don’'t need to work, 1 are the regular day. You have a budget: M, which is your paid holidays. Find the maximum consecutive day you can get with the holiday budget. For example, if M=3, you can put [0,0,0, 0,0,0,0,0,1,1], you will have 8 days of holidays.
    ai

请问lz面的是什么职位?ds还是sde?

fb的题是用dfs或dp解吗 还是我想太复杂了

面的都是DS职位

我觉得就是直接一个位置一个位置的试。但是时间复杂度是nxn (worst case). 让我改进,然后就卡住了

看不懂fb的题目…

Given A=[0,0,0,0,1,1,0,1, 1, 1]. You have a budget M (for example 2). You can consider that you have M 0s, i.e, [0,0]. Now the question is that you will put those 0s in the array A, so that you can get the maximum length of consecutive 0s. For example, if you put those two 0s in index 4 and 5, you will get A’'=[0,0,0,0,0,0,0,1,1]. So the length of 0s is 7.

I can flip M 1’‘s into 0’‘s.
I am supposed to generate maximum continuous subarray of all 0’'s

Thank you!

Case questions: 1) how to predict that a driver will cancel a ride, what kind of features you need 2) how to predict that a rider will cancel a ride, what features <= lz是用binary regression解的吗? linkedlin 的naive bayes classifier 能问下考察点是推公式?

No. This question concerns more about how you think, what kind of features are important for this prediction. For example, if a rider is young and the trip location is in downtown, she/he might have a higher chance to cancel comparing with elder rider, because young riders can walk to a bus, or subway easily.