Woofun AI reports that DeepSeek, a leading domestic manufacturer of general-purpose large models, announced on the evening of June 25 a strategic recruitment drive designed to double the size of all its departments. This expansion covers seven major categories including algorithms, R&D, operations and maintenance, products, data engineers, and functional departments, offering a total of 33 positions across Beijing and Hangzhou with internships available for every role. The initiative signals a shift from merely adding R&D personnel to constructing a complete talent pipeline spanning the entire industry chain from basic model development to front-end commercial application. Specifically, the listing includes 8 positions for full-stack development and algorithms, 5 roles in model data strategy, and 4 openings for AI core system R&D, reflecting a comprehensive structural overhaul.

DeepSeek has notably intensified its focus on agent applications within this recruitment cycle, introducing roles such as General Agent Data Product Manager for office, lifestyle, and search applications. The company defines its team mission through a specific vision: "Text has extended memory, printing has expanded the dissemination of knowledge, and the internet has extended our senses and connections. We believe that agents will represent another extension of human capabilities—breaking down the barriers of static knowledge through search, reaching into the digital world with tools, and transforming the relationship between humans and AI from dialogue to trust, thereby truly liberating human action and creativity and building a new civilization of 'human-machine symbiosis." To realize this, the firm is also seeking AI Product Operations staff focused on experience and service, alongside Professional Domain Data Product Managers specializing in small languages, medicine, and law.

Prior to this broad announcement, Cui Tianyi, head of the DeepSeek Harness team, posted a targeted job advertisement on June 21 that drew significant attention. Cui stated, "As a newly established department, the DeepSeek Harness team has ambitious goals and a heavy workload, and we are still in great need of more people. I am interviewing candidates every day and posting small ads everywhere... There are three positions available: Harness Researcher, Harness Engineer, and Harness Product Manager." He clarified that despite the increase in open roles, the recruitment process remains rigorous, typically involving one round of written tests and three rounds of interviews, with Cui personally conducting the final interview for all candidates based in Beijing.

Woofun AI data shows that the financial backing for this expansion stems from a massive financing round completed on June 16, where DeepSeek’s parent company, Hangzhou DeepSeek, raised over 50 billion yuan (about 7.4 billion US dollars), pushing its post-financing valuation beyond 50 billion US dollars (about 338 billion yuan). Several investors indicated to Jiemian News that talent acquisition and hash rate capacity were the primary drivers behind this investment decision. Founder Liang Wenfeng contributed approximately 20 billion yuan, securing his position as the largest single investor. Other participants included Tencent with about 10 billion yuan, CATL and Puquan Capital with about 5 billion yuan each, NetEase, JD.com, Monolith砺思 Capital, and IDG Capital with about 30 billion yuan each, and Zhengxin Valley Investment and Shixiang Technology with about 15 billion yuan each.

Despite the influx of capital, the company faces a complex talent landscape marked by recent departures. From the second half of 2025 to the first half of 2026, at least five core R&D members left DeepSeek to join major domestic internet companies or top AI firms. Guo Daya, a core researcher on the DeepSeek-R1 project, joined ByteDance’s Seed team; Wang Bingxuan, a key author of the first-generation large language model, moved to Tencent’s Hunyuan team; Luo Fuli, a key developer of the DeepSeek-V2 project, joined Xiaomi’s MiMo team; Wei Haoran, a core member of the OCR model development team, went to Baidu; and Ruan Chong, a key contributor to multi-modal technology, joined the autonomous driving startup Yuanrong Qixing. These exits highlight the competitive pressure on DeepSeek to retain and attract high-level expertise.

The broader market context reveals a severe supply-demand mismatch in the domestic artificial intelligence sector. According to a new economy talent report released by 脉脉, from January to April 2026, the number of AI-related positions increased by 8.7 times, and their proportion in the overall new economy job market rose from about 2.78% to 22.03%, establishing AI as a critical area in spring recruitment. Recruitment remains highly concentrated in first-tier cities and new economy hubs, with Beijing, Shanghai, Shenzhen, and Hangzhou ranking as the top four most popular cities for hiring. In Beijing specifically, the penetration rate of AI-related positions reached 30.17%, meaning that out of every 10 new positions created, 3 are related to AI.

Among the companies generating the most new positions, ByteDance took the lead, with DJI surpassing REDnote for the second time and REDnote coming in third. Major entities such as Tencent, Meituan, and Ant Group also ranked highly, while MiniMax and Zhuimi entered the top 20 list of companies with high hiring demand. Liang Wenfeng, the founder of DeepSeek, promised at an investor meeting that the company would continue developing open-source AI models and pursue the ambitious goal of achieving general artificial intelligence. He clearly stated that his main focus is on technological advancement rather than profit generation, a stance that necessitates accelerated recruitment for a company that has traditionally kept a low profile.

Technological milestones underpin this aggressive hiring strategy. On April 24, DeepSeek officially released the V4 series of large models and made them fully open-source. These products utilize a hybrid expert architecture with 1.6 trillion parameters and natively support million-level super-long context inputs, with multiple evaluation indicators showing they are on par with leading closed-source large models abroad. On May 22, DeepSeek announced a permanent price reduction for its API services: the pricing for the V4-Pro series was reduced by a quarter, and the cost for cached hits dropped from 1 yuan per million tokens to 0.025 yuan per million tokens. The prices for missed inputs and outputs were also reduced from 12 yuan and 24 yuan to 3 yuan and 6 yuan respectively. This marks a decisive pivot where capital deployment and talent acquisition converge to support a low-margin, high-volume open-source strategy.