Traditional technical interviews are struggling to keep up with LLM-assisted candidates. Rather than banning AI tools, the post proposes three interview strategies that embrace LLM usage while revealing genuine expertise: deep-dive questioning about past projects, reading a complex codebase with a deliberately limited LLM, and reviewing an intentionally confusing AI-generated code change. Evaluation criteria focus on technical judgment, problem decomposition, domain knowledge, and communication — assessing whether candidates treat LLM output critically rather than accepting it at face value.
Sort: