OM1 Interview Questions
& Process

I interviewed for a Data Scientist role in Boston. It started with a recruiter call to go over my background and paperwork. Then, I had a video call with the Chief Data Scientist, which was a bit more technical. He asked about my past work and about ML concepts, like explaining the ROC curve. If that went well, I had 4 technical interviews, each an hour long. The third was a programming session in pandas, where I had to type out functions for data manipulation like handling NaNs, grouping, adding columns based on string length, and a Python question to swap dictionary keys and values. The fourth was a discussion-based technical interview. The problem involved combining messy lab data with inconsistent labels (like 'Cholesterol in blood' vs 'chol level'). I had to design a prediction algorithm to suggest standard codes for these labels with probabilities, explaining string manipulation, feature engineering, and classifier choice. The fifth technical interview was another discussion. The task was to design an ML algorithm to predict if a patient would suddenly get very sick, based on years of vital statistics, knowing that about 1% of patients experience this. I needed to explain how to build a predictor to identify these individuals months in advance. The sixth and final interview was again with the Chief Data Scientist. This one involved logical riddles and thought experiments, like comparing multiple logistic regressions to a single linear regression for predicting pain levels, and analyzing ROC curves, including how splitting a dataset would affect AUC. They always started interviews asking for a story about a major project I handled. After all this, they said they'd get back to me, but I'm still waiting six months later, which feels pretty disrespectful.

It started with a recruiter call, then I spoke with the hiring manager about the job and my interest. After that, I had 5 interviews, each about 30 minutes. One was a live coding test with two parts: SQL and programming. The SQL part was about basic joins and aggregate functions with group bys, not too hard if you've used SQL a lot. The programming part was also straightforward if you know how to code. Another interview focused on database design, like how to create tables and schemas that connect well and are efficient. This could be tough without much SQL experience. Then there was a system design interview where they asked me to design a real-world system. Another part was a project presentation, where I talked about a past work project. They said not to stress too much about preparing it, but I'd recommend putting in effort if you want the job. I reused some material from an old presentation but tailored it. The final interview was about role-specific questions. Each interview had at least 2-3 people. Overall, it felt like a fair and well-planned process.

I had a phone screen, then a tech interview focusing on Python and SQL, and after that, I got scheduled for 5 more interviews. These later interviews delved deeper into Python/SQL, my previous projects, and posed some general 'What would you do' type questions. While I felt things went well and I was really keen on the role, despite being told 'someone will reach out shortly,' I've been completely ghosted by the company. It seems like I didn't get the job, but not even a rejection email after 7 interviews is pretty unprofessional.

The process was really involved. First, I spoke with a recruiter, then the hiring manager. After that, I had two interviews with team members, followed by a separate design exercise with them. Then, I had two more individual interviews with different team members. The final stage was a presentation of a past project to the team, with a Q&A. After all that, they ghosted me and never gave me any update. I've heard from others in the industry that this is a common experience here; they go through many interviews and exercises, only to be completely ignored.