Samsung Research Institute Bangalore Interview Experience for Internship

Last Updated : 12 Aug, 2024

Coding Round:

We initially had a coding round on Samsung's proprietary software, which only runs on Windows. The eligibility criteria were a CGPA of 7.5 or higher, specifically for circuital branches.

The coding round consisted of a single question, with a time duration of 3 hours. To clear the round, we needed to pass all 50 test cases. The question involved a matrix of size n×mn×m filled with 0s and 1s. The task was to flip the values in the columns exactly kk times. This meant selecting a column (e.g., column 1) and toggling all 0s to 1s and vice versa. The goal was to determine the maximum number of rows that could be made entirely 1s.

I successfully solved the problem by either converting the matrix rows to strings or using decimal representation through bitwise operations. The key was to convert only similar numbers or strings, checking how many operations were required (let's say k′k′). This was feasible only if k′<kk′<k and k−k′k−k′ was divisible by 2. I then calculated the maximum number of possible rows that could be converted.

Following the coding round, around 50 candidates were shortlisted for interviews.

Technical Round:

There was only one technical interview, and the order of interviews was based on CGPA, with higher CGPA candidates being called first. For those interviewed early, the interview lasted about 15-20 minutes, focusing on projects and basic data structures and algorithms (DSA) questions, such as the 3-sum problem or how to compute A+BA+B without using the direct addition operator (only using increment operators).

My interview lasted 35 minutes. The interviewer first delved deeply into my first project, followed by questions about my second project. He then inquired about my coursework, which led to discussions on linear algebra, data compression, PCA, and eigenvectors.

In terms of DSA, he asked me about the 4-sum problem. Another question he posed was about compressing a very large string in such a way that the original string could be reconstructed with good probability by randomly and uniformly selecting characters from the compressed string. I suggested using character frequency to determine the contents of the compressed string. However, the interviewer seemed unsatisfied with this response, and I struggled to provide a more comprehensive answer. The interview concluded shortly after.

Results:

Ultimately, around 30 candidates were selected, but I was not among them.