Interview 1: Googlyness
Asked some scenario based questions and- Tell me about a time...
Was able to communicate well. In my opinion didn’t use any bad example or any bad words. Could answer nicely on scenarios based question. But in some answers communication wasn’t too well. :p
Interview 2: [ 1 easy, 1 hard problem ]
For the first problem, provided solution too quick. 1 in (n^2), another in NlogN. Interviewer didn’t want me to code. Said it was a warmup. Then gave a harder version of that question on a matrix. Took time to find a soln. Then provided a bad soln of n^3. Then gave another n^2logN soln which was good. But only 15 mins was left. He wanted me to code. But coding was long. I coded 70% perfectly and talked while coding. Told that if we had time what I'd do. Asked if he understood everything. He understood and said no need to worry about other method implementation. Asked questions from the code, and suggested an improvement.
He was happy I believe.
Interview 3: [1 easy, 1 medium problem]
Interviewer gave me an easy problem. Gave him solution quickly. Asked me to code. Coded that nicely. He asked question from the implementation. I improved the solution and answered everything he asked. He asked if I know how to do a part in a particular way. Also Said It's okay if I don’t remember. I coded that too.
Then he gave me another version of the problem. I coded that too. Found some bugs myself, took a little time to fix that and then fixed it. Showed how it works in different testcases. He then changed some constraints and I answered his questions and said how would I change my code for that.
One thing is- that was an nlogn solution. He didn’t ask about improving. I had O(N) soln in mind. Forgot to talk about that :p
Couldn’t answer some small questions regarding a C++ built-in function. But I still guess that was okay with the interview.
- After interview found out missed 2 corner cases on the solution. :(
Interview 4: [ 1 Medium ]
That was my best interview. Although interviewer's communication wasn’t so clear and he was slow. He asked me a dfs/bfs type question where dfs/bfs wasn't even needed. I first explained how to do that using dfs/bfs. Then started coding in bfs. Then found bfs is not even needed. Explained it. He was happy. Coded that nicely. Asked if he has other question or should I test now.
Went through testing, Found out an error myself. Explained why It's happening and then fixed the code. Again tested and explained. Asked several times if he understood everything. He was clear. He asked some other questions about code. The whole thing was done in 30 mins.
I guess he was very happy.
Interview 5: [1 medium to hard and lot's of follow up]
The last interview was pretty challenging. He gave me a medium to hard level binary search problem. The coding part was done in 25 mins.
He then talked about CPU optimization. Asked how would I use the previous code if I had M CPU's? What would be the final time complexity. I couldn’t answer much on that. But he cleared his questions differently. It was similar to system design discussion and I wasn’t prepared for SD. I think i didn’t very well on that part, but also wasn’t too bad. He Wanted to know how I'm thinking. Whole discussion lasted till 55mins.
I guess the interviewer was also happy as he answered al my question pretty nicely. (But not sure though as the last part didn’t go well)
So, that was everything.
Conclusion:
- I found 2 problems same from LC
- Other 2 was not in LC but if someone knows DS and Algo's well, he can solve it.
- Try learning about library function we regularly use.
Update: Received offer.