The duty author | Utada
| Tiger Sniff Technology Group
The cover is from Visual China
This should be the first time in the world that "artificial intelligence and machine learning engineers' salaries are declining".
6 years ago, as a huge outlet, "artificial intelligence" spread from Silicon Valley to the world, in addition to creating several vertical industries and tens of billions of unicorns, but also created a salary myth in the field of science and technology .
After all, in 2016, Kai-Fu Lee publicly revealed: "In Silicon Valley, doctoral students who do deep learning can get an annual salary of 2 million to 3 million US dollars after graduation, and the three major companies (Google, Facebook and Microsoft) are digging people at unreasonable prices." ”
In the same year (2016), in China, many investors said in an interview with Tiger Sniff: "The four little dragons give an annual salary of 3 million, and they can't dig up artificial intelligence experts."
First, forcing digitalization and virtualizing the office at scale will only increase the urgent need for a variety of technical talents. The average skilled salary of respondents increased by 6.9% between 2020 and 2021, more than 6 figures (US dollars), setting a record since Dice wrote the salary report.
In other words, tech industry talent is not worried about finding jobs.
However, in the subdivision technology category, "artificial intelligence", "natural language processing" and "machine learning", which have set new high wage records several times in 2018, 2019 and 2020, average salaries have begun to ease, and even slightly declined - down by 2.1%, 7.8% and 8.9% respectively.
"A lot of factors are contributing to the decline in the 'value' of these skills, and one of the important things is that an increasingly large group of technicians is mastering this skill." Citing Dice's chief operating officer, the IEEE noted that over time, an increased talent pool will lead to a continued decline in pay as skill sets become easier to find.
Image from IEEE Spectrum
Of course, you might argue that this is just a sign peculiar to the United States, where AI talent is relatively abundant. For continents with large population bases and large traditional and emerging technology industries, the talent gap is still very serious.
For example, the "Artificial Intelligence Talent Report" released by the Internet recruitment platform Lagou.com in 2021 shows that the talent demand index of the artificial intelligence industry in 2021 increased by 103% compared with last year, and the average salary was 20,000 yuan, an increase of 12.4% over last year. According to Lagou recruitment estimates, the algorithm talent gap reached 1.7 million.
The data comes from Lagou.com. Since September 2021, the talent index has declined significantly
However, we need to see some "peculiarities brought by industry". That is to say, in different industries, the salaries of AI engineers may vary greatly.
For example, a number of headhunters have responded to Tiger Sniff: the salary of autonomous driving is not only high, but even continues to soar in 2021; while vertical industries with more applications such as "sound" and "text" tend to be stable and even downward.
"Algorithm talents in the field of automatic driving have risen a lot in the overall salary in 21 years, and newly graduated masters can get 500,000 to 600,000 annual salaries, but from the end of the year to now, it is tending to be stable; and some ordinary algorithm positions in other fields have actually been relatively stable since two years ago, and there will not be much increase."
Of course, a small number of scarce jobs are still rising significantly, after all, 'things are scarce and expensive'. In addition, there is never a shortage of markets at the level of the great gods. A senior technology headhunter gave feedback to Tiger Sniff.
In the past 6 years, in public reports, the artificial intelligence industry has continued to shape an image of a "serious imbalance between supply and demand", even if there is an AI bubble in 2019, a large number of startups with AI words in 2020 close, the so-called "AI algorithm talent gap" is still continuing to expand - after all, AI technology continues to sink, this trend is irreversible.
However, as Dice summarized above, a large amount dilutes the value of money.
Although Chinese and industry are in urgent need of a large number of AI-related talents, from Baidu, iFLYTEK and the fourth paradigm and many other Chinese technology companies, 5 years ago began to set up a variety of "algorithm universities and communities", cooperate with universities to cultivate talents, and even put a large number of open classes on the Internet, their efforts and appeals at the market education level are reaping rewards.
The rapid acceptance and learning and understanding of new technologies among Young Chinese people is surpassing the "massive clamor for talent shortage" in the industry from 2016 to 2019.
For example, existing open source algorithm models and open source programming tools exist in a large number of communities, including Github; now in addition to the first- and second-rate universities that must have AI labs, a large number of local vocational colleges have opened AI courses;
Even in the cram classes and online classes for primary and secondary schools now, AI has become a compulsory popular course; and around me, there is no shortage of young people and children who start programming by self-study and even brush the list...
"I have to learn programming, I have to learn AI algorithms, I can't lose at the starting line, and I also let my children brush Hacker Rank!" This is a parent's personal account of tiger sniffing, and there is a posture that when we were young, we had to learn computers and Learn English from the dolls.
Obviously, this kind of education driving force of "AI and programming for all" that has fully sunk into the education system is bound to give feedback to the industry in 3 to 5 years.
Therefore, when AI algorithm capabilities are no longer scarce, then on the one hand, it sinks faster, on the other hand, it also means that ordinary traditional enterprises can also recruit algorithm talents with normal salaries. From some domestic recruitment websites, in addition to Huawei, Ali and other large manufacturers still want to use high salaries to recruit top talents in the field, many small and medium-sized technology companies AI-related positions, the conditions are not so harsh, the salary is not high.
Image from a recruitment website
Of course, to some extent, this trend is also an important driving factor for AI startups to become less "valuable" and "noticeable".
Therefore, the myth of "artificial intelligence million annual salary" circulating in the circle five or six years ago may soon come to an end in a few years; and the continuation of this myth, yes, is semiconductor, and now even the second-rate university master's students who have just graduated are being competed for.
At the end, I would like to show you a salary list that Dice counts and ranks by subdivision technology. This list is interesting because it tells us what the most scarce technical talent in the United States really is. And half of these technologies are actually related to databases and cloud computing.
Salary rankings based on segmentation techniques. For example, Apple is also hiring the Solr team, but checked the domestic recruitment website, and the salary of such engineers is not high.