I applied online. The process took 1 week. I interviewed at Everpure in Mar 2011
Interview
Applied on-line, got 10 quiz questions in e-mail box:
Interview questions [3]
Question 1
Consider a virtual memory with M resident pages and a page reference sequence
p1, p2, . . ., pN
of N distinct requests. Assume that physical memory is initially filled with unrelated pages, N = 2M, and a FIFO page replacement algorithm is used. If the access pattern is p1, p2, . . ., pN repeated three times, then the number of page faults is:
a. N/2
b. N
c. N + 3
d. 2N
e. 3N
Consider a queue between the two processes indicated below. N is the capacity (maximum length) of the queue; e, f, and b are semaphores. "P" refers to the operation of acquiring (decrementing) a semaphore, and "V" refers to the operation of releasing (incrementing) a semaphore.
init() { e = N; f = 0; b = 1; queue = EMPTY; }
process1() {
for(;;) {
P(e); P(b); queue.enqueue(...); V(b); V(f);
}
}
process2() {
for(;;) {
P(f); P(b); ... = queue.dequeue(); V(b); V(e);
}
}
Which of the following statements is (are) true? (Zero or more may be correct.)
a. The purpose of semaphore f is to ensure that dequeue is not executed on an empty queue.
b. The purpose of semaphore e is to ensure that deadlock does not occur.
c. The purpose of semaphore b is to provide mutual exclusion for queue operations.
d. None of the above.
If V is a vector with n items and subroutine Exchange swaps its arguments, then the following code fragment sorts V in descending order.
for( j = 0; j < n - 1; j++ )
for( k = 0; k < n - j - 1; k++ )
if( V[k] < V[k+1] )
Exchange( V[k], V[k+1] );
How many calls to Exchange are made if initially, V[i]=i, for i = 0, 1, 2, ..., n - 1 ?
a. n-1
b. n
c. n(n-1)/2
d. (n-1)(n-2)
e. n(n-1)
So far had the coding test on HackerRank which was relatively easy. Two coding challenges and then 8 multiple-choice questions I think. The coding tasks are the leet-code level simple problems like counting objects in the input data etc.
Interview questions [1]
Question 1
Inferring information about the output of async tasks composed of atomic operations.
It was the first round interview and was a Hacker rank with Easy-med leetcode qs including MCQs and fill in the blanks. 2 coding qs one was easy palindrome check.
I applied through other source. I interviewed at Everpure (Bengaluru) in Aug 2025
Interview
Algorithm Round or Optical Illusion Puzzle?
Had an "interesting" experience with the so-called algorithm round. Still not sure if they were testing problem-solving skills or just hoping candidates would get lost in the formatting.
The highlight was a question on a bitbuddy tree (yep, that's what they called it) disguised in a 2D array format. Looked like a scene from Inception at first glance—layers within layers. Turns out, all it required was a plain old integer division. The challenge was more in deciphering what they were even asking, not solving the problem itself.
Would’ve appreciated a bit more clarity on what kind of "algorithm" knowledge they expect. Feels like they were going for clever, but ended up closer to cryptic.