I applied through a recruiter. The process took 3 weeks.

The process lasted 3 weeks and was divided into 3 stages. Before each stage, they sent a PDF with the description of each interview, including objectives and preparation tips. It is a tiring process, but everyone was super friendly, respectful, and made an effort to make the process lighter.

The first stage is an interview with the recruiter to describe the company and the position in general terms, understand your profile, and identify culture fit.

The second stage has two 60-minute interviews with a Data Scientist or an MLE. One of them is a Basic Programming Case in which you are asked to solve 4 HackerRank problems, sharing your screen with the interviewer and explaining your reasoning. They are easy problems, and I was able to finish in 35 minutes. The other is an Architecture Case, in which a business problem of theirs is presented (trying to reduce the number of messages that reach the support team) and you are guided with questions to build a machine learning solution. It does not involve code, but it involves decisions related to the entire machine learning lifecycle: which model you would use, how you would serve the model predictions, how you would build datasets, which metrics you would use to evaluate the model, how you would monitor if the model is working well or not, how you would fix wrong predictions, etc.

The third stage has 3 interviews. One of them is the MLE Programming Case, similar to the Basic Programming Case, but with an ML context and a small codebase related to an ML experiment workflow. The environment comes prepared in HackerRank, and there are 4 questions that involve discussing the presented problem, identifying issues with the current solution, and implementing some simple features. The most important part is to understand the problem, so ask questions, describe your reasoning, and write simple, readable code. At the end of the 3rd question, the interviewer said that I wouldn't need to do the 4th question because he already had the answers he needed and there was little interview time left, so the remaining time was used for me to ask questions. The second interview is the Hiring Manager Interview with the hiring manager. The questions are about your past experiences and challenges. It involved both technical questions and questions related to task prioritization, stakeholder relationships, etc. The last interview is the Behavioral Interview, a steroid version of the first interview, with a focus on previous experiences and cultural fit. Some objectives are to see how you deal with different situations and to see if you have self-reflection and adaptability.

Why work at Nubank?

Do you know of any initiatives to increase diversity at Nubank? What do you think about that?

In English, tell me about a piece of feedback you received. After the response, did this feedback help you in any way?

Tell me about a project of yours that failed.

Tell me about a situation where you went above and beyond what was asked of you.

What is your relationship with stakeholders in your current job?

Tell about a situation where project priorities changed midway through.

What metric would you use to compare if the features are different?