Manus replaces coding interviews with hands-on AI collaboration tests

At a Web Summit fireside chat reported on November 12, 2026, Tao Zhang, co-founder and chief product officer of the AI agent company Manus, said the firm scrapped its traditional coding test in March of this year. Candidates for engineering roles now get a four-hour practical exercise in which they use Manus to solve real problems, and hiring decisions are made by reviewing the full course of that work rather than a code sample in isolation. In effect, the company said it is evaluating candidates on their ability to work alongside AI rather than on raw coding fluency.

Role boundaries inside Manus are shifting as well. Zhang said that on the product and marketing teams he leads, product managers already submit changes directly to the code repository, and designers have started writing code too. Team members no longer split work strictly along job titles. The public careers page nonetheless still lists conventional openings such as front-end engineer and back-end engineer. Zhang described those titles as a concession to the outside world, which, in his words, has not kept pace with what teams actually do day to day.

The chat also touched on broader changes in tech hiring. Citing a recent report, Zhang pointed to U.S. software engineer hiring falling below half of its 2022 peak, with demand for front-end and mobile developer roles nearly halved, while demand for FDE positions grew 4.6x year over year. The material does not cover reactions from other companies, and the changes described apply to Manus's own recruiting process rather than to the wider industry.