GeeksforGeeks App
Open App
Browser
Continue

Adobe Interview Experience | Set 55 (On-Campus Full Time for MTS profile)

Adobe came in our college for MTS (Member of Technical Staff profile).

Round 1:

This was an online round conducted on Hackerrank platform.It had 3 coding questions.:-

1. You need to buy cupcakes as you visit from one shop to another.You can either buy a cupcake, sell a cupcake or do nothing.You will be visiting Y shops.And initially you have X cupcakes.At the end of the journey you should have Z cupcakes.Input is three numbers X, Y, Z.You need to output the maximum number of cupcakes that one can have in the journey.
2.  It was maximum contiguous sum sub-array problem.But it was not directly stated.A story was formed around it.If you observe the question correctly you will get that it maps to the  maximum contiguous sum sub-array problem.
3. You are given two strings S1 and S2. S2 contains a ‘*’ in middle not at start or end.Now the ‘*’ can be replaced by a NULL character or any other substring from s1.You need to return the number of strings from s2 that are substrigs of s1. Ex-

s1 “aaabbaaabb”

s2 “aa*b”

Ans – 12(Count every occurrence of “aa” matching with a following “b” and sum them all you will get 12)

Round 2:

Results were out after 3 days or so.20 students were selected after this round.My first question ran partial and second full.Last I attempted using find() and substr() but due to a stupid mistake, it didn’t pass.But I got shortlisted though they were also students whose q1 and q2 ran fully. Lucky me ;)!

Round 2 started with him asking me “Tell me about yourself”.He didn’t even took my resume.Then he asked me about what all projects I had done.So he took one of my project related to Naive text classification.And asked what if the search bar you have created need to have suggestions coming up once the user enters a word. Like if my db has “data Analysis” and “data mining” and “data classification”, if user types in data, he should get Analysis, mining and classification as suggestion.I gave a Trie based approach.Wrote the class for Trie. Explained the node structure, working and all. Next he asked what if your text is huge and you have lack if storage how would you store.I told him sir like we do in our computers we zip it, similarly we will encrypt our data and when needed we ll decrypt it.He asked how? I suggested Huffman Coding and explained it.Next he asked SQL, I said sir not one of my strong points.So he moved on to next.Then he asked about dynamic memory allocation, where it’s done and all, some thread related questions.Like multiple threads are running and some are not even used but user forgets de-allocate them? what to do? I said we will deallocate them gave a timestamp based approach where current time could be compared with timestamp of the thread to decide on deallocating it.He asked about problem of finding mutual friends in a facebook kind of network. what ds to use? I said graph based and would do bfs traversal and take a boolean marker to mark the friend as mutual.Also marking visited notes.

Round 3:

In second round, 7 students were rejected.

Round two started with my project.It was much deep into how you did it, tell your complete implementation process frontend, backend etc.The machine learning project was again up for discussion.Sir asked about how will you handle access to the disk when data is so large.I provided a cache based approach.Further discussion went around how will you decide what to store or remove from cache.He asked about designing a game- chess firstly but I don’t play chess which I told him.So then it was discussion on designing Snakes and ladder game.He asked about what problems will you face at every level – front end, persistent memory and in rendering if data is huge and how will you solve them.I answered as huge results may be shown through pagination.Ranking of the results based on relevance metrics(popularity of the result) can also be done.For the backend I suggested if single node can’t handle computation, we can distribute it to multiple nodes.Suggested Master- Slave Architecture.He asked what if the master fails.I said chose the best slave to be the master.Discussion went on how you ll chose it what will you do for time when there’s no master.Back-up data could be used.This round was very different as they were not asking direct question but instead checking your approach in solving a problem and once you will give one, they will add one more problem to it to make it more complex and see how you handle that again.

Round 4: