Particle.news

Anthropic Manager Ends Coding Interviews for His Team

Citing Claude’s ability to beat humans on small programming tests, he now looks for short public prototypes and portfolios that show product judgment and drive.

Overview

  • Felix Rieseberg, an engineering manager at Anthropic, said on Tuesday that he has removed timed coding interviews from his team’s hiring process because they no longer provide useful signal.
  • Rieseberg argues that Anthropic’s Claude and similar coding models can reliably solve the small, timed problems that those interviews used to test, so those tasks no longer distinguish candidates.
  • He highlights an internal example where two newer engineers built a graphical interface for Claude Code over a weekend and shipped a public version about eight days later as evidence that AI narrows traditional skill gaps.
  • For now Rieseberg is scanning resumes and portfolios for small, opinionated prototypes that show ‘hunger,’ taste, empathy and principled thinking, but he has not adopted a formal, objective replacement.
  • Reporters and analysts warn that portfolio- and prototype-based hiring raises fairness and subjectivity risks, and other big firms are only just piloting AI-assisted interview methods, so hiring practices remain in flux.