DeepSeek launches large-scale recruitment, planning to expand the size of all departments by at least double
On the evening of June 25th, it was reported that DeepSeek, a leading domestic general large - model manufacturer, released a recruitment notice, stating that it plans to at least double the size of all departments. Jiemian News checked relevant information and found that this recruitment covers seven major categories including algorithm, R & D, operation and maintenance, product, data engineers, and functional departments, with a total of 33 positions open. The work locations include Beijing and Hangzhou, and all positions accept interns.
Judging from the positions announced this time, DeepSeek is not simply supplementing a small number of R & D personnel, but building a complete talent echelon to fill the entire industrial chain from the development of underlying basic models to front - end commercial applications. Among them, there are 8 positions being recruited for full - stack development and algorithm - related positions, 5 positions are offered in the field of model data strategy, and 4 positions are being recruited for AI core system R & D.
In this recruitment, DeepSeek has increased the recruitment of talents in the field of agent applications, including general Agent data product managers (office/life/search). In the job description, DeepSeek defines the team's mission as follows: "Writing extended memory, the printing press extended knowledge dissemination, and the Internet extended senses and connections. We believe that agents will become another extension of human beings - breaking the walls of static knowledge through search, reaching the boundaries of the digital world with tools, transforming the relationship between humans and AI from dialogue to trust, truly liberating human action and creativity, and building a new civilization of 'human - machine symbiosis'." For this purpose, DeepSeek is also recruiting positions such as AI product operation (experience and service direction) and professional field data product managers (for disciplines such as small languages, medicine, and law).
In addition to this recruitment, on June 21st, Cui Tianyi, the head of the DeepSeek Harness team, posted a recruitment poster for an Agent position, which attracted a lot of attention. Cui posted that "as a newly established department, the DeepSeek Harness group has lofty goals and heavy workloads, and is still in great need of people. I am interviewing candidates every day and posting small ads everywhere... There are a total of three positions: Harness researcher, Harness engineer, and Harness product manager." According to him, although the number of job vacancies has increased, the recruitment process is the same as that of other groups in DeepSeek, usually consisting of one written test and three rounds of interviews, and he himself conducts the "final interview". The work location is uniformly in Beijing.
Regarding the approximately 50 billion yuan in financing recently completed by DeepSeek, many investors also told Jiemian News that talent and computing power may be the main considerations for the financing intention.
On June 16th, it was reported that Hangzhou DeepSeek, the parent company of DeepSeek, completed its first round of external financing, raising a total of more than 50 billion yuan (approximately 7.4 billion US dollars), and its post - investment valuation exceeded 50 billion US dollars (approximately 338 billion yuan). In this round of financing, founder Liang Wenfeng may contribute approximately 20 billion yuan, making him the largest single contributor. In addition, Tencent contributed approximately 10 billion yuan; the CATL system contributed approximately 5 billion yuan, including CATL and Puquan Capital; NetEase, JD.com, Monolith Capital, and IDG Capital each contributed approximately 3 billion yuan; Zhengxin Valley Investment and Shixiang Technology each contributed approximately 1.5 billion yuan.
From the second half of 2025 to the first half of 2026, at least 5 core R & D members of DeepSeek confirmed their departures and joined domestic Internet giants or leading AI companies. Specifically, Guo Daya, a core researcher of DeepSeek - R1, left to join ByteDance's Seed; Wang Bingxuan, the core author of the first - generation large language model, joined Tencent's Hunyuan; Luo Fuli, a key developer of DeepSeek - V2, joined Xiaomi's MiMo; Wei Haoran, a core R & D personnel of the OCR series models, joined Baidu; and Ruan Chong, a core contributor to multimodal achievements, joined the intelligent driving startup Yuanrong Qixing.
Currently, there is a gap between the talent supply in the domestic artificial intelligence track and the expansion needs of enterprises. According to the New Economy Talent Report released by Maimai, from January to April 2026, the number of AI - related positions increased by 8.7 times, and the proportion of AI positions in the entire new economy increased from approximately 2.78% to 22.03%. AI has become an important battlefield in the spring recruitment. The talent recruitment in the new economy industry is still highly concentrated in first - tier and new economy cities, with Beijing, Shanghai, Shenzhen, and Hangzhou ranking among the top four hot recruitment cities. Among them, the AI penetration rate of newly posted positions in Beijing is 30.17%, ranking first, which means that for every 10 newly posted positions, 3 are AI - related positions.
In the ranking of companies with the most newly posted positions, ByteDance leads, DJI surpasses Xiaohongshu for the first time and ranks second, and Xiaohongshu ranks third. Large companies such as Tencent, Meituan, and Ant Group also rank very high. At the same time, MiniMax and Dreame have also entered the top 20 hot - recruitment enterprises.
It is reported that Liang Wenfeng, the founder of DeepSeek, promised at an investor meeting that he will continue to develop open - source AI models while pursuing the more ambitious goal of achieving general artificial intelligence. It is said that Liang Wenfeng clearly stated that his main goal is to promote technological upgrading rather than monetization. Against this background, the always - low - key DeepSeek has to accelerate its recruitment.
On April 24th, DeepSeek officially released its V4 series of large models and made them fully open - source. The product uses a 1.6 - trillion - parameter mixed - expert architecture and natively supports million - level ultra - long context input. The evaluation data shows that many indicators are comparable to those of leading overseas closed - source large models. On May 22nd, DeepSeek issued a permanent API price cut announcement. The full - line billing standard of V4 - Pro was permanently reduced to one - quarter of the original price. The input price for cache hits dropped from 1 yuan per million tokens to 0.025 yuan per million tokens, and the unit prices for non - hit inputs and outputs were also reduced from 12 yuan and 24 yuan to 3 yuan and 6 yuan respectively.
This article is from Jiemian News, author: Peng Peng. It is published by 36Kr with authorization.