One phone screen and one on-site technical interview. Phone screen was on feature interpretation of machine learning models. On-site was 5 hours, with technical and theory questions. The technical were really nice because they gave a real data set to play with. This is a key difference from other companies and gave me a sense that they understand what data scientists do. Whiteboarding with data, in my opinion, is just nonsense. You have to see and visualize the data to understand it and what issues you may have with it. The theory questions were fairly standard stats/machine learning questions, though some were tricky. I left the interview with a good impression of the company and with the feeling that they understand the role of data scientists within the company.

Their recruiter contacted me. We talked over the phone briefly and they scheduled a call with one of the team members. The second call was ok. We spent lots of time convincing each other about our understanding of some ML techniques. After one day, recruiter sent me rejection message.