# Bharti SoftBank (Hike) Interview Experience | Set 2

**Round 1 (Written Test ) **

Duration – 90 mins

Section 1 ( Technical objective questions)

– 25 mcq’s mainly focusing on c , c++ , os , data structures , algorithms

Section 2 ( 2 Coding questions )

– Code was supposed to be written on paper

Question 1: Given a string of words return all words which have their reverse present in the string as ( (word1 , reverseword1 ) , (word2 ,reverseword2) )

eg . Input - Sachin tendulkar is the best tseb eth nihcaS Output - { ( best , tseb ) , ( the , eth) , (Sachin , nihcaS) }

Question 2: Finding the just smaller number formed using same number of digits.

Print -1 if not possible .

eg input - 371 output - 317 input - 456 output - -1

**Roiund 2 (Interview)**

First round of interview was based on basic data structures

1. Interviewer asked me to write code for the 2nd maximum element of an array . Then the 3rd and kth largest element in array was discussed .

2. Concepts related to binary search tree , uses and comparing complexity .

3. Asked about Hashing , Collisions , Implementation for hashing .

**Round 3( Interview)**

Question 1: Asked to write code for level order traversal in binary tree. Question was further modified to do a spiral order traversal. Both were supposed to be done in O(n) time complexity .

Then he asked me a puzzle

http://www.programmerinterview.com/index.php/puzzles/3-ants-on-a-triangle-riddle/

There was a small discussion about my projects then .

Then he started asking questions about operating system

– Mutual exclusion and semaphores .

– Write code for producer consumer problem using semaphores .

After I wrote some pseudo code problem was made more complex by adding further constraints .

The interviewer wanted to test my understanding about semaphores and asked some tricky and confusing questions .

**Round 4(Interview )**

Question 1: Write code to get maximum and second maximum element of a stack. The given function should be in O(1) complexity .

I gave a solution using 2 additional stacks

Then he made it more interesting by extending the question to Find kth largest number from stack at any instance in O(1) time. Stack supports push ,pop , peek , and kthmaximum function .

He was very particular about the O(1) constraint .

I gave plenty of solutions but he wanted me to reach a O(1) solution.

Finally i gave him a solution which used an additional heap and a binary search tree and was able to get O(1) time complexity .

Question 2: Given a biotonic array ( first numbers increase and then decrease ) write code to search a given number. Time complexity O(logn)

The question was further modified to write code for 4 cases to search –

1 Array could be sorted ascending

2 Array could be sorted descending

3 Array could be first increasing and then decreasing

4 Array could be first decreasing and then increasing

The idea was to first detect which case it was and then search accordingly

Time complexity O(logn)

Question was further modified to handle duplicates. In case of duplicates my algorithm became O(n)

**Round 5 (Telephonic Interview)**

This round was a telephonic interview with the CTO of the company .

First he asked me about database indexing. What ? Why to use indexing ? how is it implemented ?

I gave him a complete explanation

This is a very good explanation

http://www.programmerinterview.com/index.php/database-sql/what-is-an-index/

The he started with a small discussion about my projects and research paper .

He then moved on to data structures and asked me how to choose a relevant data structure for a given problem .

– I gave him advantages of every data structure and certain problems where specific data structures could be useful .

Then he gave a question and asked me for what data structure should be used keeping time complexity in mind

Question: Given a list of students and their marks write a function that would print all students with marks in a given range l , r

eg . Akshay 30 Atul 25 Angay 20 Sahil 10 Then if input is l = 12 and r = 26 Output - Angay, Atul

I gave a solution by indexing marks with a link list of students creating an array of link lists

He then modified the questions by adding marks in floating points were also allowed .

I gave him a solution using map (c++) based on a key value pair of marks and list of students

But he asked to optimize the time complexity

I then gave a modification of my first solution and adding concept of buckets and binary search .

He asked me a few hr questions in the end .

If you like GeeksforGeeks and would like to contribute, you can also write an article and mail your article to contribute@geeksforgeeks.org. See your article appearing on the GeeksforGeeks main page and help other Geeks.

## Please

Loginto comment...