Google plans a pilot program that uses AI assistants, specifically Google Gemini, during selected software engineering interviews for junior to mid-level roles.

The pilot targets junior to mid-career developers and emphasizes evaluating AI-use proficiency, such as prompt writing, output validation, debugging, and code reading.

In code comprehension segments, candidates will use AI to read code, identify bugs, fix issues, and optimize performance while clearly explaining their thinking.

For junior candidates, one standard technical interview round will be replaced with a broader engineering challenge and a 'human-led, AI-assisted' framework that includes an assessment of project design discussions and leadership potential.

The interview overhaul reflects a broader shift toward AI-enabled engineering practices and real-world team workflows in the AI era.

The initial rollout will begin in the second half of the year, starting with Google Cloud and the platforms and devices unit in the United States, with potential expansion to additional regions and teams if successful.

The interview format pivots away from traditional technical rounds toward open-ended challenges that emphasize creative problem-solving.

The pilot aims to mirror modern development where AI is integrated into coding and problem-solving in real-world environments.

Pilot implementations will span multiple orgs, including Cloud, platforms, and devices units, and will roll out later this year.

This move aligns Google with a broader industry trend, as AI-assisted interviews become more common with other companies experimenting in this space.

Other players, including Canva and Cognition, are already testing AI-enabled interview formats, with Canva using an AI-assisted coding round and Cognition focusing on job-like tasks performed with AI tools.

Meta has already been piloting a 60-minute AI-assisted coding test since October, using tools like ChatGPT, Claude, and Gemini.