TikTok owner ByteDance has launched a recruitment drive for senior specialists in embodied intelligence, signalling its growing ambition in the development of humanoid robots.

Volcano Engine, the Beijing-based tech unicorn’s cloud computing division, is offering a monthly salary of 95,000 yuan to 120,000 yuan (US$13,328 to US$16,838), according to a job posting.

It marked the company’s clearest move into humanoid robotics as it specifically sought humanoid specialists rather than general robotics experts.

According to listings on Chinese hiring websites, the role involved conducting research on operational algorithms, pre-training large language models and integrating hardware for humanoid robots.

Candidates were required to hold a master’s or doctoral degree in computer science, automation, artificial intelligence and other related disciplines.

Other Chinese AI giants, including Alibaba Group Holding, are also investing in humanoid robotics, which is increasingly seen as a critical arena for applying AI. Alibaba owns the Post.

Wang Xingxing, founder of Chinese humanoid robotics maker Unitree Robotics, said in Shanghai on Wednesday that this field was “one to three years away from its own ChatGPT moment” – a breakthrough after which a humanoid robot could perform 80 per cent of tasks through voice or text commands.

Wang said he expected Unitree to achieve a breakthrough in the next one to two years and to lead the global industry.

Volcano Engine’s systems could be geared towards factory applications. The company has signed a deal with a subsidiary of Huawei Technologies-backed electric vehicle maker Seres Group to jointly develop embodied intelligence and robots.

According to a statement from Seres, the collaboration would focus on “intelligent robot decision-making, control and human-machine augmentation technologies for multimodal cloud-edge coordination”.

Volcano Engine would contribute its large language models and multimodal AI, while Seres would apply its manufacturing knowledge to embed AI into physical hardware.

ByteDance in July debuted GR-3, a large-scale vision-language-action model that the company said could enable robots to carry out household tasks like hanging clothes and cleaning tables.

ByteDance’s push into robotics – and its shift from general-purpose robots to humanoids – aligns with priorities laid out by company co-founder Zhang Yiming, who is based in Singapore. Zhang regularly travels to Beijing, where some of the company’s core AI researchers are located.