I got a take home exam on HackerRank that I did not do great in. I did not perfectly solve every problem. Then it was another problem with "AI assistance" except the AI seemed programmed to be as useless as possible, it's genuinely better to just not use it. Somehow I got the next interview anyways, who also asked me some leetcode style questions and asked my to explain my experience and then next day they rejected me.

The process started with a brief 30-minute phone screening with a recruiter, focusing on my background and resume. After that, I was invite to a technical video interview with a senior engineer. This round lasted about 60 minutes, starting with a brief introduction, followed by a live coding challenge on a shared editor, and ended with a few questions about system design basics and my past projects.

The interview process was straightforward with no surprises — three coding rounds (LeetCode medium difficulty), a system design round, and a cultural fit conversation. The interviewers were pleasant and the overall atmosphere was positive. That said, it's worth noting that this format feels dated. Even before the rise of AI, LeetCode-style assessments were a questionable proxy for real-world engineering ability and cultural fit. In today's environment, where AI can solve most of these problems instantly, continuing to use this framework raises the question of what signal it's actually measuring.