I applied through a recruiter. I interviewed at NVIDIA
Interview
I was contacted by a recruiter that liked my profile on LinkedIn. The interview process then consisted of a call with a team that was interested in my skills, followed by a 5-hour panel interview divided into 2 days.
Interview questions [1]
Question 1
1. Why is batch normalization important?
2. Explain BERT, warm-up rate.
3. BERT vs GPT.
4. How to detect the horizon in a self-driving car environment.
5. CNN: how does the filter size affect the performance, depth-wise vs row-wise convolutions.
6. Explain quantization in ML.
7. Coding questions.
I applied through an employee referral. I interviewed at NVIDIA in Dec 2019
Interview
The interviewer called me and ask many many random questions diverging from my resume. The engineer didn't seem to prepare for the interview at all. The entire 1 hour have 0 coding question which is very different from any other company for new grad interview. It was straight up question related to different part including many different fields. Python, securtiy, git. The questions are very weird and doesn't look to have a focus at all. I got questions like which version of python you use for internship, after I answered python3, the interviewer was like why python 3 and not python2? I'm not exactly sure what is the expectation here.
I applied online. The process took 4 weeks. I interviewed at NVIDIA (San Jose, CA) in Feb 2018
Interview
Applied online. 2 phone screen rounds. Invited for onsite. 6 technical rounds(45 min each). Interview was focused on C++ concepts, OOPS desgin, multithreading, api design and algorithms. Didn't do well in 2 rounds. The team aren't that friendly either. The manager was even interviewing during the lunch, didn't expect that. The recruiter got back to me in a week with the result.
Interview questions [8]
Question 1
Phone round 1 : C++ concepts and a simple algorithm