For start-up companies and large companies in the AI ​​field, it is important to be able to acquire top AI talent and have the ability to deliver talents continuously. At the level of grabbing people, startups are facing a lot of pressure.
Due to algorithmic breakthroughs, policy support, and capital crowding, the artificial intelligence industry is ushered in a step-by-step development in 2017, and AI startups are constantly emerging. However, the scarcity of professional talents has become a bottleneck that hinders the further development of AI startups.
At the beginning of December 2017, Tencent Research Institute and BOSS directly released the “2017 Global Artificial Intelligence Talent White Paper†showed that there are about 300,000 talents in the global AI field, and the market demand is in the order of millions. There are more than 300 colleges and universities with all the fields of artificial intelligence research in the world. The number of graduates in the AI ​​field is about 20,000 each year, which is far from meeting the market demand for talents.
In this extremely imbalanced supply and demand, the talents of startups and giant companies are increasingly fierce. Employers are recruiting talents and even paying a million dollars a year. For startups, how to develop their own advantages and absorb more AI talents?
Talent shortage under the imperfect training system"Compared with O2O and startup companies in the sharing economy, artificial intelligence startups will have different emphasis on talent considerations," Gao Tianqi, director of Lenovo Star Investment, told reporters.
Consumer Internet entrepreneurship is "a hero does not ask for the source." Talents can start a business as long as they have insight into the product and think about the business model. However, the technical threshold of AI entrepreneurship is relatively high, and the source of talents will be relatively concentrated, mostly from top institutions in artificial intelligence and computer science.
There are not many institutions that set up AI professionals to systematically train talents on a global scale. In the decade or 20 years ago, artificial intelligence was even a very unpopular profession. Even if someone learns the direction of AI, most of them choose to change career after graduation.
“This is because AI technology does not reach practical indicators before 2013, so many AI-related students will switch to search and recommendation after graduation, and there are very few people in AI-related industries such as visual and voice. It has also led to the fact that it is difficult to directly find AI talents who already have work experience.†Shen Jian, CEO of Shen Jian Technology, told reporters.
He added, but since 2013, colleges and universities have begun to pay attention to the cultivation of AI talents. It takes several years from training to output. In the future, as the number of graduates increases, I believe that the problem will gradually improve.
In addition to the inadequacy of the university's training system, Yao Wei believes that another major reason for the lack of AI talent is the relatively weak foundation of China's related industries. For example, in China, there are only a handful of architects who have done a complete processor chip. But in the United States, NVIDIA, Intel and other companies, you can find a lot of talent in this area. China’s entry into the industry is relatively late, and the pool that can dig talents is relatively shallow.
“The scarcity of talent is a common problem for AI startups, but I believe that in the next two or three years, there will be a large number of AI talents, and people will get better and better.†Huihua Huiying co-founder and COO Guo Na The talent shortage is optimistic.
She said that just like when the Apple iOS ecosystem first came out, only very few engineers could do the App on the iOS platform. It may be that in just one or two years, engineers who understand iOS will be lacking. However, with the development of the apple ecosystem, a large number of people will quickly learn the technology in this area and achieve a balance between supply and demand of talents.
Which type of AI talent is the most scarce? Shunwei Capital Executive Director Meng Xing told reporters that the AI ​​talent team is divided into three batches. The first group is the strongest algorithmic scientists who do their own framework and cutting-edge research. These people are not many in the world.
The second batch of talents may not be able to create a framework, but they can adapt and improve on the more popular frameworks, and make customized adjustments for the projects. These people are gradually increasing due to continuous training.
The third batch is a talent based on the adjustment of parameters based on the existing framework. There are many such people. Many people who were not previously in the AI ​​industry can learn this through open classes or training.
Meng Xing believes that the most lacking in the AI ​​field is the innovative algorithmic scientists. They are the people who solve the fundamental problems at the top, and the academic and industrial circles are competing in a large amount. In addition, with the change of AI entrepreneurship to the application of the landing application, AI product managers who can understand the boundary of application requirements and the boundaries of technological capabilities are also receiving more and more attention.
Entrepreneur's talent battleFor start-up companies and large companies in the AI ​​field, it is important to be able to acquire top AI talent and have the ability to deliver talents continuously. From the level of grabbing people, startups are facing a lot of pressure.
Gao Tianqi exclaimed, "Not only BAT, but also big Internet companies with lots of money, such as fast hands and today's headlines, are also digging people at any cost. In this competition, some companies we vote for want to find Silicon Valley back. Talent, it will be difficult to convince others to join."
In this case, the role of the founder of the startup is even more important. "It is because it is very difficult for you to dig up so many cows at the beginning, so the founder itself is a cow. For example, the founders of Shangtang Technology, Defiance Technology, etc., can be technically independent," he said.
Meng Xing believes that the main reason for the company to grab people is to "see people to serve dishes." For example, some people only pay attention to money, and the company needs to adjust wages within the reasonableest possible range. Large companies have fixed salary levels, and startups will be more flexible; some people don’t value short-term wages, they pay more attention to long-term huge returns, and startups can make more incentives for equity and options; others pay more attention to their own The degree of discourse and respect, the startup company can also give it more room; some scientists pay more attention to the meaning of doing things, and the startup company can use the company's characteristics and concepts to attract talents.
From another dimension, talent is followed by people. Meng Xing said that if a startup has the resources of a well-known scientist, it can also bring talent based on the alumni relationship. Of course, the most important way to attract talents is to continue to make the company itself maintain efficient growth in business, valuation, etc., and let people think this is the opportunity.
Huiyi Huiying is an entrepreneurial team with well-known scientists. It is understood that the chief consultant scientist of Huiyi Huiying is Professor Xing Lei, the director of Stanford University Medical Physics Center. He is also the top medical imaging expert among Chinese people all over the world. The company's CEO, Chai Xiangfei, is a postdoctoral fellow at Stanford University. Guo Na is a graduate of Tsinghua University.
“The competition among AI companies is ultimately the competition of top talents. These talents are scarce resources all over the world.†Guo Na told reporters that the medical imaging AI field itself has typical interdisciplinary characteristics, whether it is technical team or marketing team. Both need a diverse, interdisciplinary portfolio. Talents with different knowledge and experience backgrounds, combining wisdom in different fields, can finally form a closed loop of productivity and break through the limitations of a single discipline.
In terms of talent acquisition, Huiyi Huiying used various alumni resources. In August last year, it launched the “Excellent Talents Program†with Stanford University, and sent AI talents and medical talents to Stanford for continuous interdisciplinary talent training. At the same time, Huiyi Huiying has also established joint laboratories with domestic universities such as Tsinghua University to continuously solve the problem of talent supply and training.
“From laboratory to product and clinical application, there is a long way to go. The company is the strongest place for industrialization and industrialization, so the linkage between the company and the university is more conducive to the cultivation of talents,†she said. Guo Na also introduced that the current salary of AI talents graduated from colleges and universities in the industry ranges from several hundred thousand to several million yuan. China's AI talent strength is in line with international standards, so the salary level is relatively close. For Huiyi Huiying, the company's monthly expenditure on labor costs accounts for more than 70% of total expenditure.
Chief scientist can't be metThe endorsement of top AI talents such as the chief scientist seems to have become the "standard" of the star AI company. After all, the industry position of the chief scientist largely determines the height that a startup can achieve.
"The top AI talents like the chief scientist can be met and not available, and there are only a handful of them around the world." Yang Ge, founding partner of Xingyi Capital, told reporters.
For example, Xingyi Capital recently invested in a company that is an AI chip, Jinyun Technology. One of its founders, Lu Yongqing, is a member of the Royal Academy of Engineering and a professor at Imperial College. He is also the leading chief scientist in the industry with nearly 30 years of artificial intelligence. Field technology precipitation. This really brings a lot of extra points to the project. Under the leadership of Dr. Lu and his disciple, Dr. Niu Yuyu, the company's technical capabilities can remain at the top of the industry.
However, the chief scientist is not a necessary condition for judging the project. Yang Ge said that Xingyi Capital’s project is based on the essence of business. The purpose is to find a truly valuable enterprise and systematically treat the company from the perspectives of technology, products and markets. Consider it.
Meng Xing believes that whether the chief scientist is an important indicator when considering a startup is mainly determined by what the startup is doing. If the company does the underlying technology entrepreneurship, basically the human ability determines the company's technology research and development capabilities, and the advanced nature of the technology research and development capabilities determine whether the company can survive. In such a company, the founding team is very important for the AI ​​talents such as the chief scientist.
If the company is doing vertical industry AI products, such as face monitoring in the security field and identifying certain plant features in the agricultural field, the company's competition is not a breakthrough academic research, framework theory, but the existing formed Productization and engineering of things, such as optimizing power consumption and optimizing user experience.
"For such companies, top scientists are less important, and even some of the less advanced scientists will have even greater benefits. Because they won't be entangled in whether this thing is leading, but more practical to focus on this. Can things solve industry problems," he said.
Nlw-Ecl Sm Fiber,Sm Semiconductor Lasers,Semiconductor Nanowire Lasers,Semiconductor Lasers Silicon Photonics
AcePhotonics Co.,Ltd. , https://www.acephotonics.com