Nubank Machine Learning Engineer Interview Questions
& Process

A recruiter reached out to me on LinkedIn around September 2025. We had a phone call where we discussed my experiences, and one of the questions was in English. Then there was a Leetcode round (easy/medium level), a System Design round, and finally an ML coding interview with 4 questions about an experiment problem on different app screens. Unfortunately, I couldn't answer the experiment question and didn't move forward.

The hiring process is quite long and demanding, with about eight stages, each lasting around an hour. In total, a candidate can easily spend over 8 hours throughout the process, including technical interviews, HR chats, values assessments, and manager interviews. Technical stages are heavily focused on real-time performance and solving specific programming problems. However, there's little emphasis on evaluating the candidate's history, previous experiences, or their potential to make a real impact on the company. Generally, the process seems to prioritize performance in specific interviews rather than a broader analysis of the candidate's skills and contributions throughout their career. Also, considering the technical demands and need for English fluency, the time investment might not be proportional to the perceived return, especially compared to international opportunities paid in dollars. Throughout the process, the length and format of the stages can be demotivating for many.

The process had 6 interviews and took about 7 weeks, usually it's 5 interviews. All interviews were in English except for the recruiter call, which was in Spanish. All technical interviews were done live using Hackerrank with screen sharing. First, I had a 30-minute recruiter phone screen to discuss the role and my background. A few days later, I had an interview with two hiring managers where we talked about my past projects in software engineering and data science, my experience, and expectations. I was also asked to explain in about 5 minutes how ML can help route customer support tickets. The next day, I received a PDF detailing the process and was scheduled for two more interviews about three weeks later. The Basic Programming Interview involved solving 4 Leetcode-style problems on Hackerrank while screen sharing. The first three were easy warm-ups, and the last one was a bit harder, but still Leetcode easy. The interviewer was friendly and made me feel comfortable. I could either talk through my thought process or solve it quietly and then explain. He mentioned he would guide me if I struggled. The Architecture Case involved discussing a business problem, framing it as an ML problem, and designing the necessary microservices. It was similar to the customer support ticket question, but focused on being proactive and preventing issues from reaching human agents. The interviewer was also very communicative. These two interviews were on consecutive days. Shortly after the second one, I was scheduled for the final three interviews, which happened two weeks later, also on consecutive days. In Hiring Manager Interview 2, I spoke with one of the initial hiring managers. It was similar to the first HM interview but went deeper into my projects, experience, and motivations for joining Nubank, as well as technology. The MLE Programming Case involved analyzing provided code that loaded a trained model and produced output. I was asked questions about using the model at a business level and making modifications for new features. Unit tests could be discussed but weren't mandatory. This was like pair-programming focused on applying an existing ML model, with some statistics but no deep ML theory. The interviewer was cooperative and provided hints and debugging help. This interviewer was the same as in the Basic Programming interview. The final interview was a Behavioral Interview with the recruiter. We discussed my projects, learnings, and how I've handled soft-skill situations, highlighting the importance of stakeholder interaction. My motivations for joining Nubank were also discussed. The recruiter mentioned that interviewers would debrief soon to decide on team fit and seniority level.