Rakuten Data Scientist Interview Questions
& Process

So first there was a LeetCode coding test, then the hiring manager interview to see if I was a good fit for the role and company. After that, there were several technical rounds where they tested my knowledge in statistics, machine learning, SQL, Python, and case studies. Finally, there were behavioral interviews to check on teamwork, communication, and cultural alignment.

So, the whole interview thing had two parts. First up was a technical check where I had to do some data wrangling with Python and SQL. Then, the second part was a sit-down with a group of people where we just chatted. They wanted to hear about stuff I'd done before, my experience with machine learning, and basically how I tackle problems. Also, they were checking how well I communicate and explain things.

The interview process started with a first round where candidates had to record 5 answers to standard questions and complete a Codility test with two tasks. The video interview format was convenient for scheduling, especially with time differences. Round 2 involved a presentation, providing a reference letter, and completing two additional Codility tasks. The advice is to do the Codility tests first, as failing them leads to automatic rejection, making the presentation and reference letter a waste of time. The first Codility test had simple tasks like FizzBuzz, acting as a filter. However, the second round's debugging and geometry problems (calculating overlap area of circles) were questionable; the geometry problem felt more like a test of looking up answers than coding ability, or it penalized those not recently out of a CS program. The focus on debugging and Big O, and the misalignment with the Data Scientist role, raised concerns about Rakuten's definition of the role, possibly closer to Data Engineer or general programmer.