Y-E-L-P MLE phone screen 挂经

Stratup: was late for 5 min or so, due to my connection issue.

Warmup: interviewer introduced himself for 3 - 5 min.

Question: segment a search string with a bigram dictionary
input: "I like ice coffee." dict = ["ice coffee", "New York"]
output: ["I", "like", "ice coffee"]

我的解答：c＋＋
1, split the search

您好！
本帖隐藏的内容需要积分高于 188 才可浏览
您当前积分为 0。
使用VIP即刻解锁阅读权限或查看其他获取积分的方式

游客，您好！
本帖隐藏的内容需要积分高于 188 才可浏览
您当前积分为 0。
VIP即刻解锁阅读权限 或 查看其他获取积分的方式
Unlock interview details and practice with AI
Curated Interview Questions from Top Companies

: "throughout the problem solving, you could have been more efficient in analyzing the issue. They would have liked to see you reach a solution faster by asking more questions at the outset to better understand the problem. "


希望对大家有所帮助。也请大家分析哪里出了问题，我以后继续努力。多谢!

看不到你的隐藏部分，我面了一模一样的职位，电面是一模一样的题目，当时是这个5分钟做完，然后followup 3-gram， k-gram全部implement完成，还有时间聊business questions。所以我猜他们可能真的expect能很快写完？我面试都用python，就是感觉能写的更快

匿名者 发表于 2022-1-13 15:38
看不到你的隐藏部分，我面了一模一样的职位，电面是一模一样的题目，当时是这个5分钟做完，然后followup 3- ...

多谢回复。看来我需要改变思路。我以前面试写题都是尽量按照工作要求来写，变量与方法的名字也写的很长，容易懂。之前是面国内公司居多，没啥问题，这次基本是第一次在墙外面，不懂规矩。。。

能问下为啥隐藏部分看不到么，积分还是？我并没有设置限制。。。

匿名者 发表于 2022-1-13 15:38
看不到你的隐藏部分，我面了一模一样的职位，电面是一模一样的题目，当时是这个5分钟做完，然后followup 3- ...

Stratup: was late for 5 min or so, due to my connection issue.
Warmup: interviewer introduced himself for 3 - 5 min.
Question: segment a search string with a bigram dictionary
input: "I like ice coffee." dict = ["ice coffee", "New York"]
output: ["I", "like", "ice coffee"]
我的解答：c＋＋
1, split the search string into string vectors with space. confirmed with intervewer that the separator is only one space, no tabs, etc.
2, split each phrase in the bigram dictionary with space: unordered_map<string, unordered_set<string>>: key: the first word in the bigram, value: the list of the second word in the bigram. confirmed with interviewer that the length of the phrase is alway 2, no exceptions.
3, iterative over the preprocessed search vector with index:
for (int i = 0; i < search_vector.size(); )
when i < search_vector.size() - 1 and the i and i + 1 are both matched: push the bigram into the result vector, i += 2
else: push the ith word into the result vector, ++i
Till here, about 30 min passed.
asked about time complexity, given number of words in search string as N and number of phrases in the dictionary as M:
O(max{N, M})
follow up: N-gram match (N = any):
I gave an inaccurate answer as a tree node, but interviewer just mentioned not necessary to write the code out though I did. He was trying to confirm if I was about to use Trie, and I finally agreed.
Till here, about 40 min passed. the last 5 min was for questions.
Result: FAIURE
Feedback: "throughou‍‌‍‍‍‍‌‌‌‌‍‍‌‍‍‌‌t the problem solving, you could have been more efficient in analyzing the issue. They would have liked to see you reach a solution faster by asking more questions at the outset to better understand the problem. "

楼上有人能说说想法吗？卡住在ngram上了

请问最开始有hr screening吗， 我收到第一轮hr面，感觉不是技术面。

匿名者 发表于 2022-1-13 18:38
看不到你的隐藏部分，我面了一模一样的职位，电面是一模一样的题目，当时是这个5分钟做完，然后followup 3- ...

请问在leetcode上面有对应的题目吗？多谢。