MSD Data Scientist Interview Questions
& Process

The interview process was quick since I didn't have the second interview. It was like other processes, but they don't follow up well, so I think they could be more personal.

This interview was pretty well-organized and the interviewers seemed to really listen. It started with a phone call, which was unexpected since my application was initially rejected, where they asked standard HR questions. Following that, an online interview was scheduled where I had to present my work. The team was friendly and showed a lot of interest. After waiting about a month, I was invited to their office. Before that, I had to complete a take-home assignment. The onsite part involved presenting my solution to the take-home problem and a live coding session. The people involved were engaged and focused on understanding my thought process. Both the take-home and live coding challenges were not too difficult, maybe easy to average. This last stage took about 1.5 hours.

The process was clear and they responded quickly. It was for a replacement position, not a new hire. The first interview was with HR, conducted in German, likely to assess my language skills. The questions were basic, based on my CV, and the HR person seemed unfamiliar with data science terminology, making it one of the worst HR interviews I've had; I even suspected it was a scam. Despite this, I was invited to a second interview, also with HR, but this time in English after a brief German start. They seemed inexperienced in hiring data scientists, with poor communication between HR and the data science team. I was told the hiring manager was a 'senior data scientist' and found their name on LinkedIn. The third interview was a 2-hour technical test using a third-party tool, which didn't clarify the job expectations. The final interview was a 2-hour session with the hiring manager and two senior data scientists. The hiring manager led the first part, asking questions to help me understand the role. A senior data scientist then focused on my past experience, probing into specific models and evaluation metrics. The other senior data scientist led the third part, asking about data preparation and model selection for the role, but it was too abstract for me. The hiring manager offered an unscheduled technical interview to let me demonstrate my modeling skills, sending data for a 2-3 hour modeling task. This involved cleaning, exploring, merging, and modeling three datasets within that short timeframe, which felt like a major red flag as usually only one or two datasets are provided. Ultimately, they offered the position to another candidate who presented a better data modeling solution, suggesting they prioritized specific modeling experience over general problem-solving and learning ability. I appreciate the team's respectful conduct throughout, even when I was asked to answer in German.