Particle.news

Anthropic Team Drops Traditional Coding Interviews

A lead engineer said advanced models that can complete common test problems make those interviews a poor measure of real-world judgment and teamwork.

Overview

  • Felix Rieseberg, an Anthropic engineering lead, said his team has removed traditional programming-question interviews and called them effectively obsolete for his hiring process.
  • Rieseberg argued that Claude and similar models can solve typical interview tasks, including small React challenges and low-level assembly problems, so those tests now often show a candidate's AI use rather than their day-to-day skills.
  • He warned that standard coding problems tend to measure tool use instead of workplace judgment, product sense, and collaboration, which he views as more important for on-the-job performance.
  • Rieseberg advised junior engineers to use AI to build fast, working prototypes to demonstrate impact and gain influence, summarizing the approach as 'hunger over seniority.'
  • The move reflects a wider industry shift in hiring practices that could force teams to redesign assessments, retrain interviewers, and weigh portfolios or collaborative tasks more heavily while critics caution about losing sight of core technical fundamentals.