The exam consisted of 3 tasks. The first included a bash exercise for finding lines in files, the second tested knowledge of message passing between processes with a large number of constraints and the third tested a common ETL procedure for training models. The bash and message passing algorithms were a little tricky as there was no way to easily test the code locally. Overall, the questions seemed relevant and the appropriate difficulty.
Can't go into specifics but the initial interviews focused on engineering fundamentals and leetcode type questions while the later interviews focused more on Machine Learning fundamentals and company fit.
Tell me about yourself and explain ur projects
How can we stop overfitting? A simple problem for calculating precision and recall was also asked.
Define a CNN, how it works and which parts of the entire CNN operation benefit most from GPU usage.
Probability question. Every 15 minutes, 20% chance of seeing one shooting star, what's the chance of seeing at least 1 shooting star in 60 mins.
Grid manipulation and graph traversal
1. Tell me about a time when to solved a problem?
- 3rd round, take home task. I have spent hours and completed the challenge. Hiring team didn't bother to send even a single line of feedback. Such an unprofessionalism.
What would you do if you have more time in work？
Viewing 1901 - 1910 interview questions