DeepSeek is looking to “at least double the size of every department” in a new hiring spree, according to an announcement made by the Chinese AI champion on Thursday.
The firm announced openings in 33 positions across seven major categories, including full-stack development and algorithms, artificial intelligence (AI) core system R&D, deep learning research, and model data strategy product management and engineering.
“DeepSeek’s hiring philosophy is to let newcomers take on the most core and important tasks directly,” the firm said in the announcement, promising that new employees can quickly grow into “top talent” in the industry and become a “driving force” behind the development of artificial general intelligence (AGI) – AI that reaches or exceeds human capabilities.
“Humanity now stands on the eve of AGI,” the announcement said. “Join DeepSeek to experience the development of AGI first-hand, sit in the front row of this era and witness the birth of a new epoch.”
The firm said it “never sought geniuses” and welcomes applicants as long as there is something within them that shines.
The late-night recruitment post had drawn more than 10,000 résumés by Friday afternoon, a company recruiter named Wang Ning said in response to inquiries on social media RedNote.
Key position openings highlighted by the firm in the announcement include server-side development engineer, pre-training data engineer and supercomputing cluster R&D engineer, as well as specialised-domain data product managers in areas such as non-English foreign languages, medicine and law.
The news came days after DeepSeek closed its first-ever external fundraising, which saw the firm becoming one of China’s most valuable AI start-ups.
The South China Morning Post reported last week that the Hangzhou-based firm raised around 50 billion yuan (US$7.4 billion) in a Series A round, pushing its post-investment value to about 400 billion yuan (US$58.9 billion).
The deal helped DeepSeek surpass the valuation of competitors such as Moonshot AI and Hong Kong-listed MiniMax AI, although it still trails Zhipu AI’s US$95 billion market value.
DeepSeek’s low profile founder and CEO Liang Wenfeng was the largest backer in the round, paying around 20 billion yuan out of his own pocket, ensuring that he would still steer the firm despite the addition of new external investors.
Other participants included Tencent Holdings and Contemporary Amperex Technology Limited (CATL), committing 10 billion yuan and 5 billion yuan, respectively. NetEase and JD.com each contributed around 3 billion yuan.
Cui Tianyi, head of DeepSeek’s newly established Harness team, has been actively posting recruitment notices on multiple social media platforms since May.
“Every member of DeepSeek’s Harness team is a newcomer, and the group is directly taking on one of the company’s most central and important tasks,” Cui wrote in a RedNote post Thursday night.
Cui, who like Liang is an alumnus of Zhejiang University, joined the firm in March after nearly nine years at Jane Street and a stint running his own quantitative trading firm, TSY Capital. His arrival suggests that DeepSeek intends to push into the fast-growing market for AI agents, targeting Anthropic’s Claude Code and OpenAI’s Codex.
In agentic AI, “harness” refers to the engineering layer that turns a foundation model into a tool-using agent capable of executing tasks.
DeepSeek has seen several core researchers leave amid an intensifying AI talent war. Guo Daya, a lead researcher on DeepSeek’s R1 model, joined ByteDance’s Seed team, while early researcher Wang Bingxuan moved to Tencent. Meanwhile, Luo Fuli, who was part of the team behind DeepSeek-V2, joined Xiaomi to lead MiMo, the smartphone maker’s family of large language models. -- SOUTH CHINA MORNING POST
