I had a phone screen with the recruiter and two pair programming sessions done via CoderPad.

The pair programming problems I received weren't from LeetCode, but they were very much in that style. They weren't difficult, but they had multiple steps, so completing them took a bit of time.

Feedback from the pair programming sessions leaned negative, which surprised me since I thought I did well and demonstrated that I clearly knew how to code. My coding style is to create a rough skeleton of the solution, then run the program with fixture data and make adjustments until it works, a bit like TDD.

However, the interviewers didn't seem to like that approach. The feedback largely focused on how my program had bugs when I ran it. It wasn't that the program had bugs and I couldn't get it working by the end of the interview, but that it had any bugs at all when I ran it. Fixing those bugs was seen as me "struggling" to get the program working.

If I had the opportunity to interview at Square again, I honestly have no idea how I could do better other than to code out a near-perfect solution that worked on the first try. Making it past this stage of Square's interview process seems like a complete roll of the dice.

When the goal of the pair programming session doesn't seem to be to solve the problem, and even small errors are enough to doom your candidacy, it's difficult to see the point of this exercise or how you're being evaluated.

Implement functions for the game Snake.