Onfido Data Scientist Interview Questions
& Process

I applied on the company website and overall I liked the interview process. It was really clear and I got all the info I needed to prep for every step. Unlike many companies that just send an email rejection with no feedback, I got a call from the recruiter who explained what the team liked and disliked about my profile. The feedback was helpful, and even though I'm bummed about the decision, I thought it was fair. The whole thing started with a recruiter call to chat about the job, my background, what I'm looking for, and how the interviews would go. Then there was a 1-hour technical interview with two Onfido employees, covering probability, stats, ML, and deep learning textbook stuff. If that went well, the next part was a panel interview with 5 people from Onfido, with each round lasting an hour and an hour break in between. The first two rounds involved solving a LeetCode-style coding problem each. The third round was about applied deep learning where I had to present a deep learning paper I chose and a past project, followed by practical questions on training and using deep learning. The fourth round was on ML and DL fundamentals, answering textbook questions on ML, DL, and math using a whiteboard. The final round focused on deploying and monitoring ML and DL systems in production and collaborating with ML engineers and MLOps. I did the entire process remotely.

The first part of the interview was with HR. They talked about the company, the job, and what the whole hiring process looks like. It was kind of a letdown because the interviewer talked way too fast and didn't seem very helpful as things went on. The worst part was that they tried to brush over the fact that the role and the job description didn't quite match up. The title wasn't for a lead role, but the description said leading a team was required. It didn't give me the best vibe about the company.

The process kicked off with a recruiter chat, which was the only pleasant part. Then came a straightforward, hour-long textbook-style interview covering topics like batch normalization and image descriptors (SIFT, HOG). Following that were two coding interviews and three more rounds featuring textbook questions, a paper presentation, and practical problems. Honestly, the whole thing felt excessively long and impersonal. Nobody ever asked about my personal experiences or interests, and people, except maybe the recruiter, seemed indifferent. Even after all the testing, I barely learned anything about the team or the job itself.