On January 27, the DeepSeek application, developed by Hangzhou DeepSeek Artificial Intelligence Basic Technology Research Co., Ltd., a domestic large-scale model company, topped the free APP download list of Apple's app stores in China and the United States, surpassing ChatGPT in the download list in the United States.
DeepSeek选人标准是热爱和好奇,DeepSeek团队完全本土人才
On the occasion of bidding farewell to the old and welcoming the new, the Zhejiang-made AI model #DeepSeek# (Deep Seek) that has swept the global AI circle like a whirlwind continues to make history.
According to Dongfang.com, Liang Wenfeng, the founder of Deep Quest, graduated from Zhejiang University with a bachelor's degree and a master's degree in information and electronic engineering. Since 2008, he has led a team to explore fully automated quantitative trading using machine learning and other technologies. In July 2023, DeepSeek was officially established to enter the field of general artificial intelligence, and has never raised external funds so far.
The DeepSeek team is small in size, less than 140 people, but the "talent density" is extremely high, and most of the members are fresh doctoral graduates, current students and master's students from top universities such as Tsinghua University, Peking University, and Beihang University. It is worth mentioning that the team does not have "returnees" and is completely local talent.
Liang Wenfeng adheres to the concept of "looking at ability, not experience", and "the criteria for selecting people have always been love and curiosity", which has attracted a large number of young high-potential talents, so that many young people under the age of 35 and no more than five years of work experience have a stage to fully display their talents.
In terms of team management, Liang Wenfeng adopts flat management, dilutes ranks, encourages free discussion and innovation, and members choose research directions according to their interests and form teams freely. When a potential idea emerges, the company allocates resources from the top down. Moreover, there is no upper limit on the movement of resources and personnel by members, there are no hierarchical and cross-departmental restrictions, and as long as there is an idea, it can be flexibly mobilized at any time.
There are no wizards, but there are geniuses to be built! The core members of DeepSeek are fresh graduates
DeepSeek's outstanding achievements stand in stark contrast to the size of its team. According to public reports, DeepSeek's staff size is less than 1/5 of OpenAI's, and there are about 70 R&D engineers and deep learning researchers in the company with more than 100 employees, mainly in the Beijing branch, and the remaining more than 30 people are in the Hangzhou headquarters, mostly front-end, product and business personnel.
Surprisingly, there are no inscrutable wizards in this company. Liang Wenfeng, the founder of DeepSeek, once revealed that the employees are all fresh graduates from some top universities, interns who have not graduated from PhDs 4 and 5, and some young people who have only graduated for a few years. In the field of artificial intelligence, he said, "the top 50 talents may not be in China, but maybe we can build them ourselves." ”
In Liang Wenfeng's view, in this wave of artificial intelligence, the starting point of DeepSeek is not to take the opportunity to make a profit, but to go to the forefront of technology to promote the development of the entire ecosystem. "For many years, Chinese companies have been used to other people doing technological innovations, and we have used them to monetize apps, but this is not a matter of course," he said. ”
The DeepSeek team is from top universities in China, and the DeepSeek team has less than 140 people
DeepSeek's success in these innovations is not a one-day effort, but the result of several years of "incubation" and long-term planning. Liang Wenfeng, the founder of DeepSeek, is also the founder of High-Flyer Quant, a leading quantitative private placement. Deepseek makes full use of the funds, data, and cards that it has accumulated quantitatively in High-Flyer.
Liang Wenfeng graduated from Zhejiang University with a bachelor's degree and a master's degree in information and electronic engineering. Since 2008, he has led a team to explore fully automated quantitative trading using machine learning and other technologies. In July 2023, DeepSeek was officially established to enter the field of general artificial intelligence, and has never raised external funds so far.
"There are relatively abundant cards, no financing pressure, and in the past few years, we only made models and not products, which makes DeepSeek more simple and focused than other domestic large model companies, and can make breakthroughs in engineering technology and algorithms." Executives of domestic large-scale model companies said.
"The success of DeekSeek also shows the power of young people, and in essence, the development of AI in this generation needs young minds." Said a person from a model company.
Previously, Jack Clark, former policy director of OpenAI and co-founder of Anthropic, believed that DeepSeek hired "a group of unfathomable wizards", in this regard, Liang Wenfeng also revealed in an interview with the media that there are no unfathomable wizards, all of them are graduates from top universities, interns who have not graduated from PhDs 4 and 5, and some young people who have only graduated for a few years.
It can be seen from the existing media reports that the biggest feature of the DeepSeek team is that it is a prestigious school and young, and even at the team leader level, most of them are under the age of 35. The team of less than 140 people, engineers and R&D personnel are almost all from Tsinghua University, Peking University, Sun Yat-sen University, Beijing University of Posts and Telecommunications and other top universities in China, and the working hours are not long.
A headhunter who is responsible for mining high-end scientific and technological talents in the field of large models told The Paper that the employment logic of DeepSeek is not much different from the employment logic of other companies in the field of large models, and the core labels of talents are "young and high potential", that is, the age was born in about 1998, and the work experience should not exceed five years, "smart, science and engineering, young, and less experience." ”
However, the above-mentioned headhunter also said that the essence of a large model startup is still a startup, and it is not that it does not want to recruit top overseas AI talents, but the reality is that there are not many top overseas AI talents who are willing to come back.
A DeepSeek employee, who did not want to be named, revealed to The Paper that the company's management is very flat, and the atmosphere of free communication is better. Liang Wenfeng's whereabouts are uncertain on weekdays, and most of the time everyone communicates with him online.
The employee had previously done large-scale model technology research and development in a large domestic factory, but felt that he was more like a screw in a large factory and could not create value, so he finally chose to join DeepSeek. In his opinion, DeepSeek is currently more focused on the underlying model technology.
The working atmosphere of DeepSeek is completely bottom-up, with a natural division of labor, and there is no upper limit on the transfer of cards and people by everyone. In the process of exploration, when he encounters a problem, he will invite others to discuss it. Liang Wenfeng previously said in an interview.
Lei Jun's talented girl Luo Fuli, who is poached with an annual salary of tens of millions, is from DeepSeek? High School Homeroom Teacher: She was involved in the development
With the explosion of DeepSeek, there is news on the Internet that Luo Fuli, a "post-95 AI talented woman" who was offered a sky-high annual salary by Lei Jun at the end of last year, is from DeepSeek. (Previous related reports: "Lei Jun's post-95 AI talent girl poached with an annual salary of tens of millions" Luo Fuli is from Yibin, Sichuan High School Class Teacher: She hasn't decided to go to Xiaomi yet, and she is still thinking about it). In this regard, on January 27, Luo Fuli's high school class teacher and Yibin No. 1 Middle School teacher Li Bo confirmed to the Red Star News reporter that Luo Fuli participated in the development of DeepSeek. At the same time, the Red Star News reporter entrusted Li Bo to convey the interview request to Luo Fuli, and Luo Fuli replied "It is not convenient to accept an interview for the time being".
Measured DeepSeek Olympiad Mathematics Problem Writing Essay: Clear Ideas for Problem Solving and Many Essay Details
On January 27, DeepSeek topped the free app download charts of Apple's app stores in China and the United States, surpassing ChatGPT in the download charts in the United States. The reporter measured DeepSeek to write an essay on Olympiad math questions. The idea of doing Olympiad math problems is clear, the level of intelligence is relatively high, and the problem solving process is detailed. The essay lists a lot of details, but it is a bit lacking in connotation and feelings.
Jimu News Comprehensive Chao News, Surging News, Red Star News, Yangtze Evening News
(Source: Jimu News Comprehensive)
For more exciting information, please download the "Jimu News" client in the application market, please do not reprint without authorization, welcome to provide news clues, and pay the remuneration once adopted. 24-hour hotline: 027-867777777.