Hello Geeks!
Samsung R&D Bangalore visited our campus on 18/9/2020 and the entire process was conducted online. Eligible for CSE, M&C, and all circuital branches.
Round 1: (Online Coding Round)
The test was proctored and conducted on Co-cubes platform and had 3 coding questions to be done within 70 mins. Different sets of questions were given to different candidates based on joining time. I got the following questions:
- Minimum Character to append at the end of a string to convert it into a pallindrome (3 marks)
- Number of Ways to Reach Nth stair taking atmost M steps (5 marks)
- Remove BST keys outside given range (5 marks)
The platform showed the results on only a few pretests while the main tests were hidden. So make sure to consider all edge cases beforehand. It was mentioned that preference would be given to optimized solutions(both in space and time). I solved all the 3 questions and qualified for the next round. In total, 21 candidates qualified for the next round.
Round 2: (Technical Round)
This round was taken on Skype, along-with colab-editor for writing the code. The interviewer was really friendly and asked me to explain my projects. Then he asked me in-depth questions about the projects. (mainly about machine learning concepts and practices).
This was followed by two coding questions:
- Implement Queue using two stacks
- Given an array of strings, you had to report all the strings which were prefixes of some other string in the array. (Hint: Use Tries)
I discussed my approaches and after getting the green signal from him, coded down the solutions. Be ready to face cross-questions like why a particular approach is better than another, real-life examples of data structures, etc. (Also prepare some low-level system design questions for this round).
The round went on for 40-50 mins and finally, I was selected for the next and final round.
Round 3:
This round was also taken on Skype. The interviewer was again very friendly and cool. He told me to introduce myself, followed by which, he started asking me technical questions related to Machine Learning and Deep Learning. The questions were aimed at checking your basic knowledge in the field. It is advisable to brush your concepts about the tech-stack and skills you’ve mentioned in your resume. A lot of conceptual questions were asked, some of them being:
- Difference between Machine Learning and Deep Learning
- Why does convolution layers work better in image classifications as compared to only fully connected layers?
- Different types of weight initialization techniques.
- Explain dropout and their resemblance to ensemble techniques
- Explain gradient descent, mini-batch gradient descent, stochastic gradient descent and which is more preferred under different scenarios
- Why Adam is more preferred than SGD, etc…
He also asked me some practical questions on deep learning, and asked me what choices I would make to improve the performance.
The round went on for 30-40 mins.
EOD: I was given an offer to join SRI-B for the 2 months internship position.