news

Lin Fei, born in 1998, became one of the first young people to be employed in the AI ​​boom. Lin Fei graduated this year with a master's degree in computer science. In June, she just stepped out of the ivory tower and joined a large company to work on AI algorithms with a salary of 500,000 yuan. "From our class, more and more students have been working in the AI ​​industry."

Since 2016, the first year of artificial intelligence, the AI ​​industry has gone through several rounds of reshuffles. With the help of ChatGPT, AI big model startups have once again been put in the spotlight. Internet veterans such as Yan Junjie and Wang Xiaochuan, as well as some middle and senior executives of major Internet companies and some academic scientists have joined this wave of big model startups.

In recent years, the number of investment and financing events in the domestic large language model has grown rapidly. From 2020 to 2022, the number of investment and financing events in each year did not exceed 5, while in 2023 there were 20. Judging from the known amount, in 2023, the amount of investment and financing in the field of large models in China has reached 6.427 billion yuan.

In 2023, there were more than 70 large-scale model startups in China in the first half of the year alone. With the advent of the AI ​​wave, the demand for talent among enterprises has naturally increased.

Data from Maimai’s high-level talent think tank shows that from January to June 2024, the 10 positions with the most talent shortage include large model algorithms, ChatGPT researchers, cloud computing, digital front-end engineers, etc. The talent supply-demand ratio is less than 0.5, which is equivalent to 2 positions competing for 1 talent.

PhD students in key laboratories receive a salary of 1.5 million yuan

"From last year to the beginning of this year, I have been interning in ByteDance's AI education product department, but many of my classmates have already decided on their future jobs." Computer science has always been a popular major in the talent market. Lin Fei told AI Light Years that his classmates received offers from BAT companies in the fall recruitment, but most of them chose development positions in the end.

"Development and algorithms are two common employment directions for computer majors. If you want to enter the AI ​​industry, you must choose more difficult algorithms and do research with a mentor who has experience with large models." Based on his curiosity about the new industry, Lin Fei chose the emerging AI direction during his graduate studies and published two industry papers.

This laid the foundation for Lin Fei's subsequent spring recruitment. She applied in March, had one round of HR interviews, two rounds of technical interviews, and the results came out in April. The entire recruitment process was incredibly fast. In just one month, she received offers for large model algorithm positions from ByteDance, Tencent, Baidu, Meituan and other companies.

Lin Fei chose one of the big companies after a comprehensive comparison, in large part because of the higher salary. "400,000-500,000 yuan annual salary is the average price for master's degree graduates in this industry." Lin Fei said that his classmates can get up to 700,000 yuan annual salary in other big companies.

“Pay equal attention to academic frontier exploration and implementation in practice.” This sentence also appeared in the mouth of Shen Jia of Fulie Consulting. He believes that this is the basic requirement for recruiting people in the current domestic AI industry.

Shen Jia is a headhunter in the AI ​​field. His team has been paying attention to the latest developments in big model technology, global talent distribution and mobility since August 2022, and helping domestic companies find high-end talents. When screening candidates, Shen Jia emphasized that companies will first evaluate candidates' understanding of the path trends of cutting-edge developments in the big model industry, and then consider whether they have practical experience in related projects.

"From the perspective of campus recruitment, corporate demand is concentrated in a few colleges and universities with relevant laboratories, and doctoral students in the laboratories are more popular." Shen Jia believes that domestic companies are currently in urgent need of algorithm pre-training talents, and most companies will recruit directly from core laboratories.

At the same time, the enthusiasm for knowledge and innovation in emerging industries are also important reasons why companies choose fresh graduates. Shen Jia said that young students know the basic concepts, the hands-on experience of emerging products and the industry perspectives.

"Because the market iterations in this industry are very fast, you have to have an entrepreneurial spirit to adapt to industry changes." A recent graduate who once worked at China Finance Digital Carbon told AI Light Years.

Maimai Gaopin mentioned in its "2023 Artificial Intelligence Talent Insights" report that from January to August 2023, 95.88% of positions required applicants to have a bachelor's degree or above, of which 44.17% required a master's or doctoral degree or above.

In terms of average salary, according to the report, the average monthly salary for newly created artificial intelligence jobs in 2022 was 43,817 yuan, and rose to 46,518 yuan in the first eight months of 2023, an increase of 6.16%.

However, for PhDs in core laboratories, they can get an annual salary of 1.2 million to 1.5 million."This is a common salary, which we call the price that a PhD student in a first-tier core laboratory can get."The so-called first-tier core laboratories refer to national laboratories and national key laboratories.

Shen Jia said, "Due to the scarcity of talent in the field of intelligence, these PhDs have more potential advantages than their master's degree classmates who graduated three years earlier and are doing similar work in large companies." Currently, many large Internet companies have begun to "reserve" talent from core laboratories in advance.

Shen Jia pointed out that this year's doctoral students have been "snapped up" by major companies, and now they are turning their attention to the talents who will graduate next year. This fierce competition for talent is not limited to domestic universities, but has also become the focus of competition in overseas laboratories and research institutions.

This year's Tencent "Qingyun Plan" is open to doctoral students graduating from January 2023 to December 2025, and undergraduate and master's degree graduates graduating from January 2024 to December 2025. Tencent has specially launched a talent program for the large model field, and the increase in the number of recruits is expected to exceed 50%. The first-line person in charge of Hunyuan Large Model personally participated in providing customized training programs for the selected candidates.

Similarly, ByteDance has also launched the "Jindouyun Talent Program" and the "Top Seed Talent Program." The "Jindouyun Talent Program" focuses on attracting students with outstanding academic research, competition results, practical experience, etc., while the "Top Seed Talent Program" focuses on recruiting top technical talents who are passionate about cutting-edge technology fields such as machine learning and artificial intelligence.

The first year of application has created more jobs

2024 is recognized by the industry as the first year of big model application. Enterprises are more concerned about the application and implementation scenarios of big models. These changes are also reflected in the recruitment needs of enterprises. For big model enterprises, algorithm positions have always been the top priority, but other types of recruitment positions are also increasing.

According to the "Niuke 2023 Fall Campus Recruitment White Paper" released by Niuke, the corporate demand for artificial intelligence/algorithm positions has increased the most, and the number of student applications has increased by more than 30% year-on-year.

A person in charge of the five major domestic model companies also pointed out to AI Light Year that the main goal of campus recruitment is to attract young model algorithm talents with high potential, with the proportion of fresh graduates being one-tenth. The company's current campus recruitment also focuses on recruiting talents in the fields of algorithms and product technology.

The relevant person in charge of MiniMax, a large model startup company, revealed to AI Light Years that starting this year, the company plans to double the number of campus recruitment positions, covering multiple departments and positions, including large model algorithms, search advertising promotion algorithms, model training and reasoning, model evaluation, product management, operations, marketing, front-end and back-end development, etc.

Liu Mo, who graduated from a prestigious university in Beijing, is one of the people who benefited from the trend of large-scale model applications. Currently, Liu Mo works at Dark Side of the Moon, doing product work.

“The big model industry can really accommodate a lot of things. Whether you are a liberal arts student studying sociology or a science and engineering student doing scientific research, you can find a place to display your talents in the field of AI.”Liu Mo noticed that colleagues working in non-technical positions around him have diverse professional backgrounds, and not all of them have directly relevant AI knowledge. However, most of them have some "experience working in large companies."

Liu Mo is no exception. "After graduation, I worked in product operations at a large company, but the department was too marginal, so I left after a year." Liu Mo believes that the dividends of the Internet era have passed, and we should embrace the next blue ocean. Therefore, with the halo of a prestigious school and a background in a large company, she jumped from the Internet to big modeling in her second job.

"Once big model companies start building applications, positions will become more diverse, and a large amount of product operations, growth, and commercialization will be generated." Looking back on the development of big model applications in the first half of this year, Shen Jia believes that students from different professional backgrounds can get development opportunities, "not only students of science and engineering, but also students studying market or marketing can benefit from it."

Shen Jia analyzed that after the widespread penetration of big model applications, enterprises will need to carry out large-scale marketing activities, which involve the creation, release, use, evaluation and reuse of content. At the same time, the demand for human resources will also increase significantly during the operation process.

However, compared with campus recruitment of many large-scale model startups, Internet giants usually provide more job opportunities for fresh graduates. "Large companies have titles and many positions, which are more attractive to fresh graduates." Lin Fei recalled the reason why he chose to join a large company, mainly because large companies are more accepting and friendly to fresh graduates.

Liu Mo has a similar view. She said that currently start-ups rarely recruit new people directly from campus recruitment for non-technical positions. "Companies are more inclined to discover and cultivate talents through internship programs, and ultimately provide full-time employment opportunities based on performance after the internship period."

"Not only Internet giants and start-ups, but also many traditional industrial companies with business scenarios and technology companies with special ecological niches are actively recruiting AI talents." Shen Jia, managing partner of Fulie Consulting, told "AI Light Years" that the development of AI has entered a new stage, and the arrival of the first year of application will give birth to more new jobs.

A McKinsey report on artificial intelligence shows that China's demand for AI professionals is expected to increase six times from 2022 by 2030, with a talent gap of 4 million.

Talent in the AI ​​industry is getting younger

Wang Sen will never forget that interview. Born in 2003, she will graduate from college next year. Unlike the traditional mainstream postgraduate entrance examination, Wang Sen decided to enter the AI ​​track. At the end of last year, she applied for an internship position at a startup company called FaceWall Intelligent Large Model and successfully entered the interview stage.

During the more than one-hour interview, the supervisor did not give Wang Sen positive feedback, but instead asked her questions about her answers. "There are so many details, and you are required to have a very deep understanding of a matter." Faced with the supervisor's strict "assessment", she was like a deflated balloon, "I thought I had no chance at all."

"Are you interested in the field of AI?" The supervisor's question ignited Wang Sen's enthusiasm. As a business school student, Wang Sen had no previous internship experience related to AI, but in this new field, many newcomers entered this blue ocean with passion. "Everyone is learning by doing, even experienced people are constantly trying and making mistakes."

After joining the company, Wang Sen noticed that there were a lot of young interns in the company. "In this startup, interns mostly work in non-technical positions, with a daily salary of about 230 yuan, which is not much different from that of large companies."

Similarly, the average age of Liu Mo's product team is only 26. "Most of the company's interns are born in the 2000s, and there are even young people born in 2007."

Nowadays, talents in the artificial intelligence industry tend to be younger, and this trend is also reflected in the professional choices of college entrance examination candidates.

A report by iiMedia Research pointed out that in China, nearly half of college entrance examination candidates (49.15%) prefer to choose first-tier or new first-tier cities as their place of study and career development. These regions have developed economies and strong demand for high-tech talents, which makes more students tend to choose majors related to high-tech industries when filling out their applications.

Shen Jia recalled that when he was helping his relative's college entrance examination candidates to apply for admission last week, he found that more and more domestic universities have opened artificial intelligence-related majors since around 2016.

At present, universities that offer AI-related majors are concentrated in first-class universities and double first-class universities. According to the 2024 ABC China University AI Undergraduate Major Rating Ranking, Tsinghua University, Nanjing University, University of Science and Technology of China, Shanghai Jiaotong University, Xidian University and other five universities are ranked in the top five in the country, and they are all S-level universities with the strongest AI majors in the country.

However, Shen Jia believes that education and talent cultivation is a long-term process that requires foresight and patience, and it is inevitably accompanied by a time lag effect. Just like a hot field like artificial intelligence, although it is currently highly sought after, no one can guarantee whether it will still maintain the same popularity and demand in a few years.

This year, talents in big models are in great demand in the market, and the last industry that was in urgent need of talents was the chip field. After 2018, with the support and guidance of national policies, the establishment of the National Integrated Circuit Industry Investment Fund, and the development of intelligent manufacturing, the chip industry was pushed to the forefront of history. Capital continued to invest in this long-dormant industry, and the number and scale of financing grew rapidly.

Xiong Fangfang, head of the ZeDe chip team, told AI Light Years, "Fresh graduates from 2020 to 2023 who choose to enter the chip field can earn 400,000 to 500,000 yuan. If they change jobs once, their salary may double, which is quite tempting. But if they don't have a long-term career plan, once the industry enters a reshuffle, their stability will definitely be the worst."

Now, large models have entered the center of the entire market stage. For graduates who have joined the workforce in the past two years, it is a good time, but as for the future, how should students choose their majors to enjoy the dividends brought by the times?

"Students must be forward-looking when choosing majors or making important decisions. This is the same as the importance of considering urban planning when buying a house. For students who have just been admitted to college, choosing key majors in key colleges may be more relevant. After all, betting on the trend is not reliable at all, and the trend changes all the time."Shen Jia suggested.

(At the request of the interviewees, Lin Fei, Liu Mo, and Wang Sen are pseudonyms)