J'ai postulé en ligne. J'ai passé un entretien chez NVIDIA (Londres, Angleterre) en nov. 2025
Entretien
Applied on their website, did 7 rounds of interviews. The whole process is pretty bad where one interview was cancelled and rearranged on the day of the interview. Coordinator isn't responsive. I was told during the interview that problem solving is the thing they look for but it turns out the specific experience is what matters. Even though it was clear in the second interview that certain fields isn't my speciality but they still arranged interviews to test me. I was told there will be a behavioural but there isn't. No standard process and free style interviews really. I would say it's better than doing leetcode (which is a complete waste of time). But i think they could have saved both sides' time on certain interviews.
Questions d'entretien [1]
Question 1
Leetcode, optimization, benchmarking, unit testing, system design for benchmarking
J'ai postulé en ligne. J'ai passé un entretien chez NVIDIA
Entretien
Very disappointing process. Very unorganized and disrespectful. They dont know what they're looking for and assign random interviewers without having a chat internally. The recruiter never bothered to reply to my emails prior to the interview. Talked directly to the HM, which was a standard screening round. I was explained what the role was and what the expectations were. Finally got a conversation a day before the panel interview where I was provided some choppy info on what to expect in the next round. One interviewer, not part of the hiring team, had no idea about the role or my background. Instead, kept asking about what they knew. At the end, I asked -- "is this related to the role, because based on my chat with the HM, I didnt think this was needed?". -- they said, "not really.". Other interviews were find, pretty standard. I never got an email from the recruiter or even an automated email.
Questions d'entretien [1]
Question 1
- Provide a code review for some python code. - gpu optimization strategies for some cuda code. - discussion on the building blocks from DL api