1. Call with the recruiter about Mistral, candidate motivations and some elements of behavioural questions about past roles
2. Coding interview - LeetCode style. Starts with a simple task, progressively made more difficult / with extra optimisations. Progressed to final part with optimisation, proposed a good optimisation but implemented it only partially correctly
3. System design - design a scalable AI system. Interviewer seemed quite opinionated and didn't seem to appreciate a solution he didn't pre-emptively have in mind. I had a feeling I would be failed 5 minutes into the interview.
Overall I think 1st and 2nd round interviewers were reasonable and personable, while experience with 3rd round interviewer was rather bad. So to sum up I would rate the whole process as neutral.
4
Autres retours d’entretien d’embauche pour un poste comme Senior Software Engineer chez Mistral AI
J'ai postulé via la recommandation d'un employé. J'ai passé un entretien chez Mistral AI (New York, NY) en févr. 2026
Entretien
Starts off with a hiring manager interview, then they schedule a leetcode style technical and a system design interview. Both of these are scheduled before any sort of power day. Overall the process was smooth but the tech screen was difficult
Questions d'entretien [1]
Question 1
They ask you to design a "CPU registration" system. Basically its just a 2D array where each y index represents a day, each x index represends a cluster, and each value represents how many CPUs the cluster has for that day. Three parts
Part 1: Given "schedules", produce the resulting CPU count. Schedules are in the format "X day, Y cluster number, Z number of CPUs reserved
Part 2: Greedily count the max number of CPUs you could get across all days if you only selected one cluster. Very similar to "Best time to buy and sell stock" leetcode easy
Part 3: Same as above, but if you can only switch clusters K times. Very similar to "Best time to buy and sell stock IV" DP Leetcode Hard