YOE : 8 months at a fintech startup [2021 passout | tier 2.5]
Recruiter reached over email for a short call to discuss, he told there would be total of 5 rounds (1 phone screen + 3 techincal onsite + 1 Googlyness).
Every round is only 45 mins.
After the discussion he scheduled a phone screen round with 3 weeks of prep time.
After 3 weeks:
Phone screen : recruiter was friendly and started off with a small discussion on my current role (2 mins, i kept it short).
Question 1 : easy bst question - find element in bst when bst implemented with 1.pointers and 2.arrays.
Used too much time here, I have only 20 mins left for Q2.
Question 2 : medium level graph traversal question (can be solved with bfs after a small observation -
the observation/statement is crucial for proving the solution). I had only 10 mins left for coding, coded fast,
there was a silly bug which he pointed out and I corrected it.Got a mail in 2 hours from recruiter that i passed phone screen, He informed that I did really good in the round.
Scheduled onsites with 1 week gap on 4 consecutive dates(Thursday , Friday, Monday and Tuesday).
Onsite Round 1 (3.00 PM to 3.45 PM) : Introduced ourselves and current role discussion (5 mins).Interviewer explained me the question(didn’t paste any content on screen),
we deep dived in to several examples. I started with brute force and wasted so much time there. This is an interval based question and there are some boundary cases we need to handle.
I explained the brute force solution and he asked me to optimise it.
With only 20 mins remaining i explored the optimised approach, With a little bit of observation i converted the question to a 2 pointer scenario and explained.
Now only 10 mins are remaining, I have to code and handle special/boundary cases.I felt the pressure and tried to calm myself down.
I coded a partial solution and modified it bit by bit to handle special cases. He was satisfied with the solution.
Overall i used 2 extra minutes.
At the end I asked some google culture bases questions and about his work.Onsite round 2 (2.15 PM to 3.00 PM) : Introduced ourselves and current role discussion (3 mins). The interviewer was a SE with 5+ YOE.
With 40 mins remaining, she pasted the question and I started discussing it. This was a 2d coordinate space based question , the brute force solution was a direct one.
We have to optimise the computation mainly (using prefix sums + binary search).
Basically we have to reduce the problem into 2 similar independent problems. I did a good thing of explaining my intuition and little mathematical proof of my approach, she was satisfied and asked me to code the solution.
I looked at the time it was 2.30 PM( I thought only 15 mins are remaining and started to code fast), Somehow I managed to code the entire solution in 15 mins.
(In my mind I’m thinking : shit! the round is done and I didn’t dry run any cases)
The interviewer was patiently listening to my code walkthrough and asked me to dry run a test case , I thought yeah sure but constantly stressing out because she might mention the time spill as my fault.
I dry ran a case, this involves huge computation manually so it took around 5 to 7 mins for running a small test case(we only ran half of the test case ) - In the end we got correct answer and she was satisfied!
At 2.56 she asked if I have any questions for her, I asked a couple of culture. + role based questions.
Even after the interview I’m still thinking that I used extra 15 mins , In reality I used only 45 mins. I realised it when I’m about to sleep in the night and felt really great.
```
Onsite round 3 (2.30 PM to 3.15 PM) : Introduced ourselves and current role discussion (3 mins). He was the least interactive among all interviewers and didn’t give any hints when I’m struck.
He pasted the question, After 10 mins , I explained a straight forward brute force solution and spent next 15 mins optimising it.
My solution has a bad worst time complexity- It performs best for an average scenario. I thought he would approve it but asked me to think of some other data structure to solve it.
I explored some other ways of doing it but I’m not sure about the exact data structure implementation (interval trees with balanced BST).
After some time he gave me the choice of coding my solution if I wanted to OR figure out the most optimised one.
I have two choices here
- Code my semi-optimised solution
- Figure out the algo for optimised one (no time for coding)
I went with 1 and coded the solution (and did an extra space optimisation) and he was convinced that it was a working solution.
At end he asked if I have any questions for him, I asked a couple of culture. + role based questions.
Onsite round 4 (2.00PM to 2.45PM ): This round was rescheduled and there was a weak gap. The interviewer was interactive and we had a casual discussion and some managerial questions.
———————————————————————————————————————
The recruiter asked for references and some questions to prepare my candidature for review.
It’s been a little over 2 weeks since I gave my last interview , The recruiter messaged yesterday that “the recruitment process is slow at the moment and it may take a little longer time to arrive at the final hiring decision”. He added that I did well in the interviews and after a decision from business they’ll continue the process.
————————————————————————————————————————
I’m not a fan of 45 mins rounds, a coding round should at least be of 1 hr duration.
Will update the result once I receive it.
-------------------------------------------------------------------------
Background about me : Started cp in 4th semester (there were a lot of pauses but during my 7th sem, i reached 2050 - 5* on codechef(only short contests) and 1800 on codeforces). I rarely participate in contests now (due to job).
My cp count is good - above 800(i use stopstalk to track) but that includes several coding platforms(cc + cf + atcoder + hackerrank + hackerearth) and duplicate problems.
Lc count : 99 easy + 220 medium + 64 hard - i did 60 hard ones during this prep.
CP is not mandatory but it makes `the art of figuring out` and mainly `observing patterns` easy.