Opendoor Data Scientist Interview Questions
& Process

A recruiter reached out to me for this role. I currently work as a Data Scientist at a similar company, doing modeling for products. After the initial recruiter call, I was immediately invited to a live coding interview. I asked if we could do a QA-type interview instead because I'm not very comfortable with live coding, but they proceeded with the coding interview. A senior data scientist conducted the interview, and another senior research scientist observed via Zoom. Initially, I was asked to explain ML modeling concepts in detail. I was told I had ample time, so I explained thoroughly and answered questions, making no errors as I already do similar modeling work. Then, I was given a dataset and asked to evaluate it and derive some conclusions through aggregation using Python. I started the task, but the senior data scientist said he was running out of time and asked me to explain my approach verbally instead of coding. While I don't think I said anything nonsensical, the time limit meant I don't believe he listened carefully or asked me the right questions. I thought I would at least get to another round, but I was informed the next day that I wouldn't be proceeding. It felt like they had already decided on someone else. I'm not sure if that's what happened or if the position was closed, especially since I later heard Opendoor laid off 22% of its workforce. Even though I had asked about job security and was told layoffs had already occurred, more happened again.

The interview process involved two stages. The first stage was a data challenge focused on predicting housing prices using the KNN algorithm. The second stage was a technical screen on Coderpad. I found the coding questions to be okay, specifically an anagram problem. However, due to it being my first coding screen, I became nervous and took too long on the first question, so the interviewer didn't even get to the second one. I was rejected about 30 minutes after finishing the second round.

I applied online and after a technical phone screen, I was invited for an onsite interview. The onsite consisted of three hour-long pair programming sessions that mirrored the day-to-day tasks of a Data Scientist here, and one whiteboard design discussion. I hadn't done pair programming before, and while it felt stressful initially, it ended up being a really enjoyable experience, almost like exploring a dataset and solving a modeling problem with a colleague. The interviewers and management I met were all competent and nice people.