The interview process consisted of two main stages:

• A typical call with the recruiter to align potential candidates with the role.
• A code exercise through a tool similar to Coderpad.

Before the exercise began, the interviewer and I discussed and agreed on my approach. We also agreed that we would not be looking for a performant or elegant solution.

I chose to solve the problem using Java, which resulted in a verbose and, admittedly, not very elegant solution. I was able to walk through my logic and explain my thought process as I moved along. Excluding compilation errors, I got the solution right on the first try.

After completing the exercise, I took the initiative to discuss potential code improvements and how I would have approached the problem in a real-world scenario. We also discussed the current state of technology at Affirm and their ongoing challenges, which aligned with my own past and current experiences.

The recruiter contacted me the following day to inform me that they had decided not to move forward. With little to no feedback, the reason for their decision was unclear, especially since I successfully solved the code exercise during the interview.

I suspect the interviewer expected a more elegant solution, based on comments he made about Python's utilities making the exercise easier to complete. He also suggested finding a more performant approach as I neared the end of the exercise, even though we had initially agreed not to focus on optimization. It appears the interviewer is junior; this is unsurprising, as juniors often overemphasize minor details, much like we all did when we were junior.

Given a collection of lists of co-occurring terms, return a map of each term to the terms it co-occurred with the most.