Tsinghua University smart driving leader "puts out fires" for Nvidia

Riding on the " AI " wind, NVIDIA's market value has soared to one trillion this year, and its performance has achieved triple-digit growth during the industry downturn. It is easy to laugh at the top of the mountain as its friends and businessmen struggle under heavy loads.

Those who do not plan for the eternity cannot plan for a moment.

NVIDIA founder and CEO Jen-Hsun Huang is very aware of danger in times of crisis, and "only 30 days until bankruptcy" has almost become his mantra. Naturally, we don’t want the current prosperity to be short-lived.

When the AI ​​business became a cash cow, Nvidia did not stop. Instead, it poured a lot of resources into the automotive business, viewing it as one of its future performance growth points. And it is not satisfied with only providing high -computing hardware for the market , but also has the ambition to become a smart driving solution provider.

It is precisely because of this that Wu Xinzhou, who was the vice president of autonomous driving at Xpeng Motors, was recruited to Nvidia to take over the intelligent driving software business and report directly to Huang Renxun. Is this high-profile industry leader the right person Huang Renxun is looking for? Can Nvidia's automotive business be on the right track in the future under his leadership?

All this will take time to answer.

Wu Xinzhou, a master of Tsinghua University

Wu Xinzhou is a technology giant. This is the industry consensus and one of the reasons why Huang Renxun took great pains to invite him to Nvidia.

Wu Xinzhou received a bachelor's degree in electrical engineering from Tsinghua University in 1998 and a doctorate in electrical engineering from the University of Illinois at Urbana-Champaign in 2004. Since then, he has worked for Flarion Technologies Inc. and Qualcomm , and his achievements in autonomous driving and other fields have been praised by the industry and academia.

In early 2019, Wu Xinzhou joined Xpeng as vice president of autonomous driving at the invitation of He Xiaopeng, chairman of Xpeng Motors.

Xpeng Motors' biggest label or feature is its relatively excellent autonomous driving technology capabilities, but its initial road to self-research of intelligent driving software was not smooth. But after Wu Xinzhou joined, he reorganized the technical structure of Xpeng's smart driving team and integrated the team, which allowed Xpeng to gradually achieve "a far lead" in the field of smart driving among local OEMs.

Xpeng Motors' main model P7 was launched in April 2020. At that time, Xpeng replaced the Bosch solution with a self-developed algorithm, more than a year earlier than Ideal and NIO.

With more focus, more experience and technology accumulation, Xpeng's autonomous driving system level is higher than that of most competitors. For example, the intelligent driving systems of Xpeng P7 and P5 are mainly based on the NVIDIA Xavier platform, with a computing power of only 30 TOPS. Many of the higher-priced Ideal and NIO corresponding models are based on the NVIDIA Orin platform, with a single chip having a computing power of 254 TOPS. But no one in the industry would think that the level of intelligent driving achieved by P7 and P5 would be lower than that of models such as Ideal ONE.

A former Xpeng employee told Mr. Xin, "Inside Xpeng Motors, the evaluation of Wu Xinzhou is very positive. After all, the intelligent driving led by Wu Xinzhou can be said to be Xpeng's highlight."

But can just being a technical expert drive the team to continuous success? Have you seen how many technology giants have lost their jobs to major companies in recent years?

In fact, what is little known to the public about Wu Xinzhou is that he is also a "ruthless person" at work, a manager who strives to drive the team forward.

A headhunter close to Xpeng Motors' smart driving team told Mr. Xin that whether Wu Xinzhou has a job or not and what his management style is, it is almost clear by referring to the situation of Xpeng Motors' smart driving team in recent years.

Although Wu Xinzhou looks gentle, according to 36Kr, he acts very decisively and has a full "Eastern" management style. Some employees described it as a "stress test", "If you are out in the field today, and you have to go to the Guangzhou headquarters for a meeting tomorrow, you must get it." Someone from Xiaopeng commented that Wu can reach anyone at the bottom, and the hierarchy cannot stop him at all. "As long as he pays attention, he will have meetings with people at the bottom every day."

According to Wanwan Auto, an autonomous driving engineer from a leading domestic car company told the "legend" of the collaboration between Xpeng's Chinese and American autonomous driving teams: There is a 12-hour time difference between China and the United States, but the Guangzhou team can often test the North American team when they wake up. software, "(it seems) Wu Xinzhou is online 24 hours a day."

There are reports that Xpeng internally believes that this "strong" integrated management style is the key to Wu Xinzhou's ability to push Xpeng Motors' high-end smart driving into mass production despite a chaotic background.

Crazy enrollment expansion

After joining Nvidia, Wu Xinzhou, who is vigorous and resolute, continued his work style.

In the past four months, while Wu Xinzhou was building a team, he was also making some changes in the smart driving workflow. These actions have brought visible changes to NVIDIA's original smart driving team: faster software iterations, increased vehicle testing, changes in code integration methods, etc. The most significant is the sudden increase in working hours and workload.

Nvidia has autonomous driving development teams in China, the United States and Germany, and the teams in the three places maintain the same code base. "After he (Wu Xinzhou) came, there was no situation where anyone was involved. Everyone was involved. It's just that The local team gives priority to solving local problems. "In terms of smart driving testing, NVIDIA's vehicle testing has almost never gone outside the company. Wu Xinzhou, who has repeatedly emphasized the importance of "real vehicle testing" at Xiaopeng, breaks through NVIDIA's previous model of smart driving.

However, if the Nvidia Intelligent Driving Team wants to achieve what Wu Xinzhou expects, various internal changes and potential tapping are not enough. More fresh blood must be introduced to improve the overall combat effectiveness.

On November 28, Nvidia’s official public account released recruitment information for the smart driving team. Wu Xinzhou appeared in the recruitment process as the head of Nvidia’s China autonomous driving team.

The next day, Wu Xinzhou forwarded the recruitment tweet in his circle of friends and said, "NVIDIA has hundreds of self-driving positions open in Shanghai, Beijing, and Shenzhen. Using China's experience and talents, based on decades of intensive cultivation of the AI ​​ecosystem, and Car manufacturers will work together to deliver global self-driving products and accelerate the implementation of high-end self-driving cars around the world. If you are interested, please send me a message on WeChat.”

Source: Screenshot of NVIDIA official account

According to the recruitment information, NVIDIA’s intelligent driving team has five departments open for recruitment this time, including the autonomous driving software group, the autonomous driving platform group, the system integration and testing group, the map and simulation group, and the product group. The number of positions open to each department varies, but overall, there are 25 positions in total.

Wu Xinzhou also put forward requirements for the abilities of job seekers: they must have a strong background in the professional field; strong self-motivation and the pursuit of excellence in technology and products; identify and adapt to NVIDIA's diversified work environment, and have a good relationship with colleagues from different backgrounds. Good communication willingness and ability.

It is worth mentioning that some netizens shared their experience of interviewing Nvidia’s autonomous driving team on a social platform. According to his description, although he solved all the algorithm questions in each round of interviews, he still failed to get the offer because the position was not well matched to him.

Source: Screenshot of the above netizen’s interview post

An insider close to NVIDIA's recruitment revealed that NVIDIA has opened some level 5/6 positions, benchmarking against Alibaba p8/9. In addition to being able to use English as a working language, other recruitment requirements are not much different from those of domestic first-tier manufacturers. .

The p8 and P9 levels basically belong to team sub-projects or team directors. This group of people can be said to be high-level talents, which are relatively scarce in the industry. Under many conditions, there may not be many people who can ultimately meet NVIDIA's interview requirements.

The above-mentioned people familiar with the matter said that there is no shortage of opportunities for high-level talents in the current market itself. However, NVIDIA's position in smart driving hardware is compounded by Wu Xinzhou's influence. When news of NVIDIA's smart driving team expansion comes out, most algorithm software engineers will be interested in asking about it, even if they didn't originally look at the opportunity. "But in reality it may be difficult to reach a cooperation intention. Basically, the following factors will be considered: team integration, salary and benefits, reporting relationship, 24h development (time difference between China and the United States), etc."

But for the future development of the team, we must continue to replenish high-quality combat capabilities no matter how difficult it is.

At present, outstanding smart driving talents are basically concentrated in China and the United States, and engineers in China are more accepting of Wu Xinzhou's work and management style. In a previous interview, Wu Xinzhou also said that he knew "how terrible it is for Chinese engineers to roll up."