I applied online, was given a take-home assessment, and asked for an onsite. I echo the sentiment of others interviewing for this position. The onsite consists of 3 interviews where they provide a dataset and ask you to perform classification in 1 hour each. The first two interviews are definitely modeling based (timeseries), so do something quick as mentioned by others. The last just a basic SQL/Pandas interview. The Data Scientists are typical of SF "coder bros". They act high and mighty and are eager to jump on any mistakes you make (coding including). It is difficult to model a dataset you have no idea what the features are and then with the constant reminder of the time limit. It is nice that their VP of DS provides feedback, but I would really encourage them to review their interview process. Without sounding cynical and bitter, I would avoid this company. For such a young company with that many negative experience....maybe its not us?

As other candidates have mentioned, I received an automatic e-mail saying that Granular received my application and would like to move forward with a take-home assignment. They estimated it would take two hours to complete this assignment. The assignment took about two hours to complete, but I spent a couple more improving my writeup, making nice graphs, and explaining my results. I submitted the assignment, and the next day received a generic email saying that the company will not be moving forward. It is quite obvious that this company does not know how to interview, is unclear as to what they are looking for in candidates, and that they do not understand the data science process. I would highly suggest not to waste time engaging with this company.