The cross-border influx of talents into intelligent driving: rationality and madness coexist

The topic of "intelligent driving circle entry and education" has comprehensively "invaded" major network platforms.

Author | Jelly Ping

Edit | Wen Liang

"Three years of experience in searching and promoting, is there any opportunity to transfer decision-making planning control?" The internet is so lifeless. ”

"Vehicle engineering wants to develop in the direction of driverless, intelligent vehicles, and the Internet of Vehicles, what courses and knowledge should you teach yourself during college?"

"Sincerely ask, automatic driving VS Internet factory programmers, the starting salary is almost the same, in the long run, which has more career development?"

......

Similar topics of "smart driving circle beginners" have comprehensively "invaded" Pulse, Zhihu, CDSN, and even major online platforms such as Tiger Punch and Bilibili in the past year.

Education double reduction caused online education and training to brake sharply, the Internet industry comprehensive anti-monopoly, community group buying from hot to cold, called the game version number approval..... In 2021, in the face of the disappearance of policy dividends and industry dividends, from start-ups to giant enterprises, the Internet industry has seen a wave of layoffs, which has intensified so far.

In contrast, the smart car track has a very different picture.

In recent years, the automotive industry has ushered in a critical period of transformation, intelligent, networked, electrification has been elevated to a strategic height, intelligent driving companies have emerged, the industry is extremely hungry for software algorithm talents, the main engine factory, Tier1, intelligent driving company has gradually become the foothold of programmers in large factories, fresh graduates of the Department of Electronics or Computer Science department have begun to no longer regard the Internet factory as the first choice for employment, vehicle engineering, automation and other professional traditional auto people are also thinking about how to transform into the field of intelligent driving.

Amplified anxiety and desire are driving people from all walks of life to enter the smart car and autonomous driving industry.

Transformation: some for high salaries, some for "no rolls"

Lekuo, who is a second-year graduate student of communications, is trying to take advantage of the summer vacation to find an internship for an autonomous driving algorithm position, but she is still hesitating whether to turn to the direction of decision-making, Slam, planning or control, "I don't plan to go to the volume perception, how do you in the group determine the direction?" It looks silly to yourself."

Recently, she reported Baidu Apollo's live training camp course, added a lot of automatic driving learning exchange groups, and also submitted resumes to different positions in a number of intelligent driving companies, but still felt that there was no direction, so she couldn't help but ask in an automatic driving learning exchange group.

There are many people in the group who are not from the department of electronics or computer science like Lekou, some are product managers in the new forces of car manufacturing, some are doing automatic driving sensor integration in traditional car factories, and junior students who want to know which company's intelligent driving business is more promising are helping her to give advice.

On the pulse of the workplace social platform, about "how do Internet people transform into car companies" and "The Internet encounters layoffs, can I go to car companies?" The topic was raised early, the answers and discussions were constantly updated, and hr and headhunters in the circle were also active here, and threw an olive branch to potential candidates, saying that "without car-related knowledge, you can also transform into autonomous driving."

If you look at the resumes of the current main engine factory, Tier1 or intelligent driving enterprise algorithm engineers, you will find that many of them are indeed halfway out of the house, and even some positions are directly hired by fresh graduates of machinery and vehicle engineering, and they have not been in-depth contact with intelligent driving related knowledge and projects before. Chen Jun is a typical example.

Chen Jun's undergraduate and doctoral degrees are both in physics, but this is not his ambition. Out of a strong interest in computer vision and deep learning, Chen Spent two years, starting from Python and then moving to the field of autonomous driving, and is currently the director of algorithm research and development of a new car-making force and the head of the autonomous driving regional platform.

Zheng Hao, co-founder and CTO of Zhijia Technology, has also undergone a "transformation". Before entering the smart driving industry, he founded a social platform big data analytics company, and later served as vice president of Yahoo's Global R&D Center in Beijing, where he was responsible for the development of applied scientific research and personalized platforms.

The factors that attract non-scientific talents to take root in the field of intelligent driving are high salaries.

Headhunting company Kemai Manpower found that intelligent driving practitioners are almost the highest paid group within the main engine factory, far exceeding other practitioners of the same rank.

Moreover, in addition to conventional Internet talents, the source of talents who have entered the field of smart cars is diversifying.

According to the "2021 China Autonomous Driving Technology Talent Flow Development Status Report" released by the headhunting company Yihe Human Resources Group, the relevant talents of electrical and electronic manufacturing, information technology and services, scientific research institutes, computer software and Internet companies are likely to become the reserve force of cross-border car manufacturing in the future.

According to the statistical analysis of the group's talent database, most of the fresh graduates who choose to enter the field of autonomous driving are also from majors such as mechanical engineering, vehicle engineering and communication engineering.

One of the reasons why information technology and services, computer software and Internet employees choose to enter the field of autonomous driving is that the working hours in the field of autonomous driving are more reasonable and healthy under the condition of similar salary levels.

Although the influx of intelligent driving talents come from all walks of life, from the perspective of research direction, the whereabouts of these cross-line talents often point to three types of positions, namely algorithms, simulations and tests.

Among them, the algorithm class can be subdivided into decision algorithms, SLAM algorithms, planning algorithms, control algorithms and perception algorithms, and the test direction is divided into software testing and real vehicle testing.

Because the control algorithm is directly facing the vehicle, the algorithm entry learning is relatively simple, many automotive personnel with vehicle engineering and mechanical engineering backgrounds will aim at electronic control system engineers, electronic control algorithm engineers, vehicle engineers and other positions when transforming into the field of intelligent driving.

Image source: Qingyan Car Union

The position suitable for Internet people, is concentrated in product planning, Internet of Vehicles, intelligent cockpit, L3/L4 level of automatic driving and other fields, a car company employees said that because of certain code capabilities, but also have product thinking, originally in the Internet enterprise to do product managers, user experience, interaction design, artificial intelligence or maps and other directions, you can go to the field of intelligent driving to do car machine system, interaction design, automatic driving algorithm development, etc.

Wang Yucheng, head of 3D vision of Xiaopeng Automobile, said on the pulse that candidates working in positions such as "data mining", "algorithm research and development", "engineering development" and "basic platform construction" of Internet companies are hot in the new energy industry.

Past the peak of the talent gap?

The industry is most lacking in algorithm research and development engineers.

Peng Chun has been a headhunter in the automotive circle for 13 years, and many executives of auto industry chain companies have successfully transformed and successfully jumped jobs through her hands, but at the moment she is struggling to find senior experts who understand both algorithms and cars.

When the industry was the most frantic, a typical scenario was that when the company finally found the right algorithm engineer, it found that the candidate was holding seven or eight offers in his hand, and one salary was higher than the other.

"In fact, this candidate may not be particularly good in every way, and the frequent job hopping has pushed the salary upwards."

Peng Chun believes that the algorithmic position has become impetuous to a frightening degree, "If you pay high, I will be higher than you." Candidates are also starting to be impetuous, 300,000 this year, 600,000 next year, and 900,000 or even 1.2 million the year after."

Enterprises are also involuntary, there have been human resources director to Peng Chun poured bitter water, both understand the algorithm and understand the car people are too few, "not a will be difficult to seek, but a person is difficult to seek", in order to attract talents, enterprises can only keep increasing prices.

In addition to high salaries, because the autonomous driving industry as a few years ago was still in the demo stage, the requirements for enterprises to recruit talent in the algorithm will be much looser, and candidates can easily find a good job as long as they master some skills related to automatic driving.

More and more people are fighting in the field of intelligent driving, and institutions that provide corresponding online courses have also emerged in recent years, such as Deep Blue Academy, July Online, AutoMotive Academy, etc. Of course, many colleges and universities have begun to build autonomous driving-related training camps for autobots with vehicle engineering backgrounds, and even claim that in just 5 days, they can lead students to complete the minimum subset of autonomous driving, learn the entire research and development process, and create an unmanned vehicle.

However, compared with the barbaric growth period a few years ago, Peng Chun believes that the current autonomous driving industry has passed the peak of the talent gap.

In the past few years, whether it is a car company or tier1, it is adding an intelligent driving department or setting up an intelligent network-related subsidiary and a forward-looking research institute, a large number of traditional automakers have completed the internal transformation and transfer, and intelligent driving companies have also supplemented low-level R & D personnel through a variety of ways.

"In fact, the industry's demand for talents is far from saturated, but now the source of talents is much more, the pool has become larger, and the market has cultivated a group of professionals who understand both cars and algorithms, so companies are not as crazy as they were in previous years."

These factors have also contributed to the increase in the entry threshold to a certain extent.

For example, Li Xianggen, co-founder of Yun Whale Intelligence, once said in an online salon that due to the recent improvement of the industry's landing requirements and the collapse of a number of companies affected by the epidemic, while the number of posts has been reduced, the existing companies have higher requirements for job seekers, "I hope that we can screen out excellent talents who can face practical problems, coding, solving problems, and landing." And these talents are usually rare in the pool of social recruitment resumes. ”

Liu Langechuan, director of algorithm research and development at Xiaopeng Automobile, also revealed that many candidates' project experience is often broad but not deep, and their resumes are very eye-catching but they do not seek much understanding of details.

"Compared to the large and comprehensive project experience, at Xiaopeng we prefer candidates with a solid foundation and high potential. During the interview process, our self-driving team does not take too difficult algorithmic questions and pays more attention to basic and conceptual questions. It is imperative not to make quick gains. ”

The supply of senior talents in the intelligent driving industry is still in a serious state of short supply, and the action of enterprises to recruit each other has not slowed down.

Gao Xiang, head of the Zhixinger positioning group, said, "Some graduates of professional laboratories are not worried about finding jobs, basically they are poached after graduation, and there is no employment pressure." ”

Not long ago, because the level of recommended candidates did not meet the requirements, Peng Chun was also intimidated by the HR of an enterprise.

"HR said, 'Do you think companies are willing to spend millions to train him?'" We need a mature, ready-to-use one, not someone who hasn't been in the position for a day and let them get promoted. ’”

Peng Chun also revealed that compared with buying technology, the phenomenon of digging people in China's autonomous driving industry is very common, and she can't pick up orders, "Sometimes in order to dig up suitable people, even expert dogs, we will manage."

"Experience becomes a shackle"

Even if the transition is smooth, the relevant job offers, cross-industry talents and experts who are poached need to go through a period of adjustment, but not everyone can successfully adapt to the new job.

A headhunter in an autonomous driving circle shared the phenomenon of candidates jumping from the Internet of Vehicles to traditional car companies in the pulse: some car companies will indicate the Internet background in the recruitment position, but when Internet talents are hired, there may be many cultural adaptations, and finally choose to leave.

The same story is played out in some Internet-based automotive software and emerging car companies, "especially those that advertise themselves as technology-based innovative companies, only recruit talents with Internet company backgrounds, and disdain product technology experts with car company backgrounds."

In recent years, more and more new cars and intelligent driving functions are being rapidly launched and iterated according to the rhythm of the Internet, many traditional auto people are afraid of this, while Internet people believe that traditional car companies follow the old ways, do not understand innovation, product forms urgently need to change, and the debate between the two sides is endless.

In Peng Chun's view, people from Internet companies and traditional car companies represent two completely different corporate cultures and work models.

Internet companies are accustomed to small steps and fast iteration of products, have a certain tolerance for failure and error, mostly adopt a flexible working system, and the product thinking of traditional car companies is to try not to make mistakes, and because of the many parts involved, all parties need to cooperate with each other, forming a uniform working time, most of which are nine to five.

When people with two very different habits of mind and work collide with each other, it is inevitable that the problem of corporate organizational culture will arise.

For example, Peng Chun revealed that due to the in-depth cooperation with a large Internet manufacturer, the relevant personnel of the large factory settled in the headquarters of the car company, which made the pressure on employees increase suddenly, "the work is very volumetic, HR is still interviewing until 11 pm."

In contrast, the division of the car company still retains a relatively relaxed working atmosphere. "Both of them (traditional car companies and Internet companies) are tolerating each other and trying to find a suitable coexistence model, but this requires a balance and coordination at the top."

In addition to the need to run into each other's work rhythms, the integration of talents from different fields must also consider each other's product development thinking.

Si Chen went from the MCU development post of the traditional car company Tier1 to an Internet company as a senior engineer of smart cockpit.

He posted about the shocks of switching from the traditional automotive industry to the field of intelligent driving:

Before joining a new company, I thought that Internet companies entering the automotive industry would try to establish processes and capabilities for the development of automotive software. When I joined the company, I realized that I was very wrong, and my previous work experience with Tire1 was completely useless. These experiences have become shackles, making it difficult for me to adapt to the development process and rhythm of Internet companies. So far, I haven't adapted to this rhythm.

Si Chen believes that compared with the traditional car Tier1, Internet companies have stronger software capabilities and tool development capabilities, as well as a more open attitude to the use of open source technology.

"Most of the tools used in the traditional tire1 are provided by specialized companies, such as Vector and EB. And the Internet company I work for has a dedicated tool development team, and most of the applications I test for software are developed by that team, which is a really great ability."

Karin Xianren is a senior programmer who has been engaged in back-end research and development in Internet companies for more than ten years, and five years ago he turned to the field of self-driving car infrastructure research and development.

In his view, because the development, testing and operation environment of the Internet industry is very mature today, developers of Internet companies usually only need to pay attention to the operating system interface layer, without going deep into the hardware layer.

However, when developing automatic driving systems, because the software system is closely coupled with the vehicle platform, vehicle-side hardware and sensors, and the industry is in a stage of rapid evolution, there are many models and sensor models, technical standards and interfaces are changeable, and developers need to understand all levels and links from software to hardware to the system.

"It is best for developers to have the opportunity to personally participate in the whole process of hardware configuration and software deployment and debugging, so as to deepen the overall understanding and avoid blind people touching the elephant."

Therefore, as intelligent driving technology moves towards product landing and functional car, compound talents with cross-industry and interdisciplinary backgrounds are being sought after by enterprises from all sides of the automotive industry.

Kemai Manpower said that it can integrate the traditional direction with emerging technologies, and talents with different technical backgrounds are highly sought after in the market, and with the gradual Internetization of the automotive industry, the phenomenon of "inner volume" in the industry will become more and more serious, and "big and small weeks" will gradually become standard.

However, Peng Chun believes that although in the long run, the demand for talents in the field of high-end automatic driving will become more and more large, but in the next five years, because autonomous driving technology can only be closed roads or commercial vehicles landed, there will also be a number of intelligent driving companies to fall, the industry will need less and less talent.

"Compared with a few years ago, people who are now transforming into the field of intelligent driving across the industry will get inferior opportunities or benefits, and when looking for opportunities, they need to adjust their mentality."

(Le Kou and Chen are pseudonyms)

END

Intelligent driving "investment" accelerates the inner volume: pick and counter pick

"Unmanned" driving is encountering a "employment" shortage