I had Google interview and feedbacks wasn't good enough to pass HC. My feedback package didn't even go to HC even though my behavioral and system design feedbacks were good. I was asked one LC hard question during one of the interview and an open ended question during another interview (given 100K points within coordinate system, find the coordinates of smallest square that contains cluster of at least 15% of all points). I prepared very hard for the interviews and feel very heartbroken. If I probably study another month, I would probably fail again from these questions. Is it normal to get this level of coding questions? I think Google should improve its process to send each candidate's packet to HC after onsite interview. After all HC exists to normalize the process. They could provide feedback to interviewers and they could remove such extremely difficult questions from Q database. HC has no clue such questions are being asked during some interviews such as mine b/c my feedback packet didn't even go to HC. TC? Peanuts #google
Interview questions are luck and timing friend.. don't take it personally, i have a Google offer , but got easier questions at Microsoft but couldn't crack it
How long did you take to prepare before your first interview?
Wow! Never heard of Microsoft interview difficulty > Google interview difficulty... guess it really does come down to luck, huh?
So how do you even solve that
I could be wrong, but a greedy approach works. Say a tuple coor is 10 bytes, so you have a Mb of data and can justify preprocessing. From there, I want to turn the problem into trapping rainwater problem: sort, treat x as i in trapping rain, treat y as height in trapping rainwater, and use two pointer sliding window looking for 15k points in your window. Time wise, you incurred a factor of nlogn for the sort and a factor of n for the window, but no new data I don’t think. Might be a bucket sort approach if X or Y is bound by 100k.
I mean the brute force approach would be n^3 straight forward iterate over all points, three nested loops, the innermost loop counts the points contained in the rect created from outer 2 loops. If count/size >= 0.15, then return rect. I am still thinking about a more optimized approach, though. Not 100% convinced with snap's answer yet, but might as well be correct.
To answer your question. Keep trying. Algorithm lottery is not fake. Many people here got in after multiple attempts. Each interviewer have different set of favorite questions. Not all are hard.
Why don't more people reject interview feedback on HC level? If I was on HC and saw a question that is purely math with No hire, I would seek to re-do the interview.
I got to Google on my 5th attempt after 12 years of trying, how crazy it is!? When I got an offer I didn't event want to consider it, but then decided to join and I am happy I did it.
Dang. Hard as hell. Don’t feel bad. Just a shitty question that shouldn’t be given in an interview with such time constraints
I had a similar experience. Asked about interval trees and the interviewer was very cryptic about this too and refused to acknowledge any other approaches I went with. Not really something covered in LC. It does suck but you have to move on. I think a lot of this comes down to luck. You can't ever know everything. Just gotta try next year again if you really want to get in
Chill u’ll end up doing crud no matter what and where
cheer up MS announced a pay bump recently!
Same , I was asked a very shitty question : Not posting the exact same question but rather similar version of what I saw : Given a set of mines and each of their coverage radius find a path from src to destination without hitting a mines coverage. The path is doesn’t have to be made of straight lines and can have all weird shapes . I felt it was too hard for me to solve , but then also saw some folks producing answers on leetcode . Wonder how they did that ?! The levels interviews have fallen to these days 🤦🏽♂️ no real software engineering, just some complex math that you’d never use at your work
Looks like a minor modification of https://www.codechef.com/problems/MAXGRID, a hard problem in a long challenge, ie., a coding contest where you solve 10 problems over 10 days. Definitely not something one should ask in a coding interview, looks like you had a bad interviewer.
How does one asking questions like these even become an interview in the first place? Seems like someone wants to feed off their ego by asking unreasonable questions
google don’t has any “interview review process”? People ask shity questions in Meta will be forced to take additional training, or even be removed from the pool if they don’t improve.
I am going to search this Q in our Q database
Same
You people have a Q database?