ThoughtExchange Machine Learning Engineer interview questions

The interview process is composed of 1 standard hr interview, 2 technical interviews with the team and a final interview with a manager (eventually a ceo). The hr was very professional and efficient. And this is all about that company. For people that have background in machine/deep learning and looking to work on that, do not waste time applying to them. Although they advertise in their job posting for NLP and requirements for skills of state of the art models like Bert, they don’t have people that qualify for that and will test you on pure software engineering skills. First technical interview was about algorithm/data structure. Two people interviewed me, the manager and one data analyst. Both were polite. But they are both of physics background and don’t have a sense of time when coding because they expect you to solve 1 medium and 1 hard leetcode exercices in 45min. The manager expects you to walk him to the most optimized solution, what he has the solution for from internet obviously. And you need to stick to what he has as a solution so for example he asks you not to use a counter. Aftermath, I discovered that the solution in stackoverflow doesn’t contain counter which was a natural choice for the problem. Plus the test is conducted via an editing tool not even a coding tool so you can run your solution. I ve experienced interviewing for big companies like MS and google and the first round is always testing your SWE skills but it is only 1 medium/hard leetcode exercice, and you will discuss with your interviewer who will most of the time give you a hint and discuss the path to the solution with you. Here no hints, the manager will let you know that he is busy working and once you finish let him know. So basically he will check the lines, if each line of your code corresponds to the stackoverflow/ leetcode solution that he has, it is ok otherwise he will ask you about the specific line to change. Plus when you apply to NLP jobs , generally the swe exercice is about manipulating strings as this is the core of nlp, here neither of the exercises is about strings as the first one is about integers, the second one is about grids and graphs. This is basically one of the companies that think that they need NLP but don’t know what NLP is about and how to test against it.