AI Reshapes Software Engineering Hiring: From Coding Tests to AI-Assisted Decision-Making

The rise of generative-AI coding tools has forced tech firms to rethink interview formats. Some companies now ban AI during tests to curb cheating, while others permit it, creating a moving target for candidates and recruiters alike.

CNN

A Google research report found 90 % of tech workers use AI for writing or modifying code, a 14 % increase from the prior year. Yet a separate survey showed 46 % only “somewhat” trust AI-generated code and 31 % say it “slightly” improves quality. AI was cited as the top reason for job cuts in April, according to Challenger, Gray & Christmas.

CNN

Stefan Mai (former Meta/Amazon engineer, Hello Interview co-founder), Varun Mohan (director, Google DeepMind), Madhu Kurup (VP Engineering, Indeed), Boris Cherny (head of Claude Code, Anthropic), Greg Brockman (OpenAI president), David Barajas (software developer), Sujata Sridharan (ex-Bolt engineer), Jordan Leonard (COO, Leopard.FYI), plus hiring firms such as Google, Meta, Amazon, Anthropic, OpenAI, Indeed, and Challenger, Gray & Christmas.

CNN

• 90 % of tech workers use AI for coding tasks (Google report).

CNN

• Trust levels: 46 % “somewhat” trust AI code; 31 % report only slight improvement.

CNN

• AI cited as primary driver of April layoffs for the second consecutive month.

CNN

• Anthropic’s Claude Code contributed 100 % of Cherny’s product work over a 30-day span.

CNN

Varun Mohan emphasized that AI is a tool, not a replacement: developers should spend most of their time deciding what to build. Madhu Kurup likened AI to Google Maps—helpful for navigation but not for setting the destination. Stefan Mai warned that AI has “hit engineering interviewing like an atomic bomb,” reshaping assessment priorities. Greg Brockman cited an OpenAI engineer who used AI to compress a week-long system change into hours. Boris Cherny noted that his recent contributions were entirely generated by Claude Code.

CNN

Employers have instituted bans on AI assistance, screen-sharing requirements, and “no-AI” clauses to prevent cheating. Critics argue that traditional coding tests no longer measure real-world skills such as delegating tasks to AI, brainstorming, or rapid trade-off analysis. Rapid shifts in language requirements—e.g., a Ruby on Rails demand later dropped because AI can translate code—underscore the instability of current assessments.

CNN

Developer David Barajas reported multiple interviews where recruiters explicitly prohibited any AI tool, even demanding desktop monitoring. Sujata Sridharan observed that most firms still use legacy tests focused on raw code, widening the gap between interview expectations and daily AI-augmented work. Jordan Leonard described a “moving target” where language specifications change weekly as AI capabilities evolve.

CNN

Sources reveal a tension between high AI adoption (90 %) and modest trust (46 % somewhat trust, 31 % slight improvement). No data currently exists on the long-term performance of candidates who use AI during interviews, leaving a gap in evidence for best-practice hiring models.

CNN

• “I would say AI has hit engineering interviewing like an atomic bomb,” — Stefan Mai, cofounder, Hello Interview

CNN

• “We think developers should spend most of their time trying to figure out what they should build,” — Varun Mohan, director, Google DeepMind

CNN

• “The first thing they say is, you’re not supposed to be using any AI tools, no AI assistance, nothing to help you solve this problem,” — David Barajas, software developer

CNN

• “It feels like a moving target on literally a weekly (or) monthly basis,” — Jordan Leonard, COO, Leopard.FYI

CNN

• “There is that gap, it’s just grown wider (because of AI),” — Sujata Sridharan, former Bolt engineer

CNN

Companies are piloting on-site half-day problem-solving sessions, emphasizing trade-off reasoning over pure coding. Some firms now allow AI assistance during assessments, aiming to evaluate how candidates orchestrate tools rather than write code unaided. The industry’s next phase will likely standardize AI-augmented evaluation criteria to align hiring with the evolving engineer role.

CNN