After the first introduction phone call I was invited for a coding/ML interview. The interview had two parts. The first part was a coding question for about 25 minutes. The question was to implement an algorithm that samples from a discrete distribution given the probabilities of the events and I was allowed to use only basic random generators. The question was easy to solve by building intervals on the real line in proportion to the given probabilities and applying a binary search. The second part was questions about ML foundations. They asked me to explain logistic regression and its loss function and how to optimize it. Also some questions about metrics to evaluate ML algorithms and evaluation methods in general