They interview for a level (SDS) not a specific role in a team. After all 6 interviews I was non the wiser as to what area there were jobs available in which was very frustrating. This meant I couldn't gain any feel for the potential work or team environment in talking to the interviewers.
The interview process is long but well detailed: recruiter/HR call, stage 1 technical interview, then a 4 stage loop of SQL, Stakeholder, Presentation, and Case Study. Although the four stage loop is long, in total it felt similar to other interviews where they would be combined into one or two interviews.
I was repeatedly told the role was about taking a wide problem and developing it into a detailed, clear problem space yet the case study was not at all like this. Although I excelled at the other 3 loop interviews, I fell down on the case study. I was given a statistical experimentation case study (not a role I was after) because there were several roles like that availble. As this was focused on experimentation - not ML as I clearly stated I was after - I felt like my time had been wasted.
Aside from this, the interviews were well detailed in advance and the interviewers very lovely and their enthusiasm for their roles shone though. They focused more on communication than technical skill with the only coding interview basic SQL (grouping, joining, windows) and no technical questions about machine learning, or dealing with large data.
Case study: A stakeholder shows you a graph of number of restaurants vs order rate, and says "If we increase the restaurants available to customers would that lead to growth of orders?". Discuss.