The interview process involved the following stages:

• Recruiter screen
• Take-home assessment through a GitHub private repo with a Pull Request submission.
• Onsite rounds, including:
  • One pair-programming session.
  • One code review session, incorporating system design questions related to the code under review.
  • Two separate behavioral interviews with Engineering Managers.

Leveling is determined throughout the process.

Overall, the process itself seemed good, and I was very excited about it. The take-home assessment felt practical and was enjoyable to complete, even though it could easily take longer than the suggested two hours. I also appreciated the structure of combining a take-home test with pair-programming and code review sessions.

The behavioral interviews with the Engineering Managers were standard.

The real issues were as follows:

1. The live pair-programming session was very poorly set up on their side. The recruiter provided incorrect instructions for environment setup. The interviewer mentioned they had dropped Docker and were using a different setup. Consequently, I spent most of the interview time troubleshooting the environment and trying different approaches with the interviewer until "Plan B" was finally ready. This distraction impacted my performance. Furthermore, the "Plan B" environment prevented me from properly testing my code, as it relied on a public and unstable endpoint.

2. The code review session was conducted in Ruby/Rails, despite me clearly stating my language preferences multiple times. The Pull Request submitted for review lacked any context, leaving me with only ten minutes to read the Ruby on Rails code, infer the necessary context, and provide review comments. This was all within a language and opinionated framework I had no prior experience with, and the position was not specifically a Ruby role. As a result, I did not perform well.

3. Finally, the recruiter sent a follow-up email stating, "Thank you for speaking with INTERVIEWER NAME(S) about GitHub’s..."

All in all, the process from GitHub was very unprofessional and sloppy.

Questions around API endpoints, JSON object manipulation, and caching.