Cambricon posts job ad to recruit AI software talent for 11 teams to enhance ecosystem

Cambricon published a recruitment announcement via its WeChat account on September 2, 2026, seeking software talent for 11 teams. The roles span drivers, compilers, high-performance communication, computing libraries, deep learning frameworks, and ecosystem component development, as well as inference and training solutions. The initiative is designed to build a more comprehensive AI software ecosystem and enhance competitiveness through software-hardware synergy.

The announcement noted that over the past year, the proliferation of AI code generation tools has reshaped software development models, significantly improving productivity across industries. Innovative products that once required hundreds of people can now be developed, launched, and operated by just three or four. This shift in development models is driving sustained rapid growth in computing demand. Meanwhile, post-training is playing an increasingly important role in enhancing model capabilities, and model iteration cycles have shortened dramatically, placing new demands on underlying hardware and computing infrastructure. The market therefore increasingly needs AI infrastructure that offers massive scale, strong ecosystem compatibility, and high stability.

According to Cambricon's 2026 interim report, the company continued to optimize and iterate its basic system software platform in the first half of the year. On the training software front, it refined its software stack around ecosystem building, model adaptation and performance optimization, low-precision training, and tool development, achieving phased progress in large-model training adaptation, training efficiency, and system stability.