Seventeen-year-old Chen Guangyu from Shenzhen gained fame for contributing to the AI paper "Attention Residuals" and the development of the Kimi K3 model. During high school, he interned at an overseas AI company and joined the Moonshot AI team. At age 17, he received five job offers with annual salaries exceeding one million U.S. dollars, sparking a fierce talent war in the AI industry and highlighting the trend of younger, higher-paid top AI talent.

Author: Investment World

Source: Mars Finance

A high school student has gone viral in the AI community.

The incident began when 17-year-old Chen Guangyu posted a farewell message to Kimi. In just a few words, he offered no explanation for his departure or announcement of his next destination, prompting immediate speculation about where he might be going.

Chen Guangyu last gained attention in March this year. At the time, he was still in his final year of high school and co-authored the paper "Attention Residuals" as a joint first author. The paper was shared by Musk, and the related technology was later applied to Kimi K3, making the teenager from Shenzhen an instant sensation.

Around the same time, renowned Silicon Valley investor Deedy Das mentioned an unnamed 17-year-old teenager who had received five job offers, each worth over $1 million (approximately 7 million RMB). The overlap in age and background quickly led the public to think of Chen Guangyu.

17 years old, five job offers, $1 million—enough to show the magic of this era.

17-year-old teen quits job, reportedly holding a million-dollar offer

On October 15, the 17-year-old teenager from Shenzhen posted a farewell message on social media, saying that the past year had been the best of his life so far. Reflecting on this experience, he said, "Thank you to everyone, especially Du Yulun for believing in and hiring a 16-year-old me."

When he joined, Chen Guangyu was still a senior high school student. In March of this year, his name appeared alongside Zhang Yu and Su Jianlin as co-first authors on the paper "Attention Residuals." Musk subsequently shared the paper, calling it "impressive," drawing attention to this young intern on the Kimi team.

By July this year, Attention Residuals had been incorporated into Kimi K3, becoming one of the underlying technologies used in this 2.8 trillion parameter model. Chen Guangyu also contributed to the preparations leading up to K3’s release. In his farewell post, he still remembered those days working alongside colleagues while watching the World Cup.

Regarding the decision to leave, Du Yulun revealed that the two had several long conversations. Though reluctant, he still wanted Chen Guangyu to explore as much as possible at this stage of life and follow his new curiosities.

Chen Guangyu did not explain his next steps in the farewell letter, only writing at the end: “There is more to explore.”

In the comments section, new invitations follow.

Sholto Douglas, Head of Scalable Reinforcement Learning at Anthropic, mentioned that he had just met Chen Guangyu last week and half-jokingly asked, “Still planning on Carnegie Mellon? Want to come work together in San Francisco?”

Thank you, Sholto! I’m heading to the Bay Area,” replied Chen Guangyu. Some speculated this meant he would join Anthropic, but he quickly followed up with a short “no :)”.

Just days before this farewell letter was sent, a report about a $1 million annual salary offer had already made a 17-year-old boy a focal point.

Last week, Menlo Ventures investor Deedy Das shared an encounter he had with a teenager who, at age 14 and with limited English proficiency, began teaching himself AI using large models. Since then, the teenager completed a one-year internship at a top AI company, traveled halfway across the globe, and enrolled as an undergraduate at a top-ten U.S. university. Two weeks after enrolling, he decided to drop out.

According to Deedy, the teenager, who was 17 at the time, already had five job offers from top-tier AI labs in San Francisco, each worth approximately $1 million or more.

Seventeen years old, one year of internship, millions of dollars... When these pieces of information were combined, they quickly prompted questions, and some people linked this experience to Chen Guangyu.

However, Deedy did not publicly disclose the boy’s name. When pressed for details, Deedy shared the boy’s reason for wanting to go to San Francisco: everyone there is amazing and is doing what they love.

I only want to learn; I couldn't care less about money.

Hottest high school student, from Shenzhen

Chen Guangyu's story begins in Shenzhen.

Born in 2009, he grew up in Shenzhen and previously attended the Shenzhen Nanshan Campus of Peking University High School and华润小径湾贝赛思国际学校, graduating from high school this year.

Like many Gen Zers, he was deeply curious about the world. In middle school, he tried registering a Shopify store using his parents’ identity, organized an open-source community for teenagers, built crypto bots, and even dreamed of becoming a professional skier.

When preparing to apply to university, Chen Guangyu paused activities he had been doing for college applications but didn’t enjoy. He picked up a book on AI safety he had long wanted to read, then moved on to physics, the universe, and Jobs. Later, in his blog, he recalled that during those weeks, he studied purely out of curiosity, without needing to prove anything to anyone. Yet, seeing his classmates busy with application materials, he sometimes felt anxious, wondering if what he was doing mattered.

The turning point came in February last year, when Chen Guangyu participated in a high school hackathon with a project called ThirdArm—a concept for a "third mechanical assistive hand for humans." It was through this project that he met Dong Kehan, a founding team member of YC China and Jiqi Chuangtan. Dong offered him a profoundly impactful suggestion: focus on more cutting-edge technologies.

Afterward, Chen Guangyu began using AI to read research papers and track open-source projects on GitHub, gradually building his understanding of machine learning. In his view, the method of following outdated video tutorials from a decade ago was far too slow. “In this era, AI is the best teacher.”

Soon, a technical blog post brought him job opportunities. At the time, Chen Guangyu shared his insights on DeltaNet. Although his social media account had only 20 followers, the post attracted a reply from researcher Yang Songlin, and was later seen by the CEO of Silicon Valley AI startup Tilde Research, who reached out to schedule a call.

Two months later, 16-year-old Chen Guangyu received a job offer from the company, flew alone to Palo Alto, California, and began his seven-week internship.

This experience impacted him far beyond simply joining a foreign company. Dong Kehan once reminded him: Don’t think of yourself as an intern—think like a CEO: If this were your company, what would you do? That advice later became his guiding work philosophy. From that point on, his involvement gradually expanded from technical issues to more complex areas, including operations setup, fundraising timing, investor communication, and long-term strategic thinking.

The rest of the story is well known. Chen Guangyu joined Kimi to continue researching model architectures. Less than a year later, he decided to leave again to explore new possibilities.

At 17, many choices are far from final. As Chen Guangyu wrote on his personal homepage: “Explore the possibilities of intelligence.” As for where this path will lead, he is still moving forward.

Grab them before graduation

AI's wealth story has made its way onto young people's paychecks.

Domestic internet giants have offered starting salaries of up to 8 million RMB to new graduates. A professor from Tsinghua University’s Department of Computer Science revealed that PhD graduates specializing in large models can earn over 6 million RMB annually, while master’s graduates can earn over 1 million RMB. The intensity of talent competition has even extended to graduation defense events, where HR representatives from major companies arrive with flowers to celebrate students who have already been hired.

Visiting Anthropic’s official website also reveals a public reference: the research position in pre-training lists an annual salary range of $350,000 to $850,000, with specific compensation depending on the role level. These positions directly involve model architecture and large-scale training.

As for young talents who have already emerged, the offers are even more staggering. Mark Zuckerberg personally called Matt Deitke, then 24, to recruit him as an AI researcher. To secure his joining, Meta offered a final compensation package of approximately $250 million over four years, with as much as $100 million potentially paid in the first year.

The founder of a tech giant personally invited young people in their twenties—highlighting the urgent competition for talent.

Extreme pricing applies to only a few, but the competitive logic is not hard to understand: rapid model iteration, uncertain technological pathways, and the addition of key researchers may bring new directions.

The recruitment timeline has also moved up to before graduation. This year, when Tencent launched its Qingyun internship program, it actively encouraged lower-year students to apply, offering unlimited compensation and additional incentives for outstanding contributors. Core businesses such as HunYuan, Yuanbao, and WeChat are open to interns, and top performers may have the opportunity to lead projects and work directly with technical leaders.

Bytedance Seed highlights young researchers’ achievements on its careers page: one individual contributed to the pre-training architecture of Seedance 2.0 less than a year after graduation; another, while still an intern, already had their work deployed in DouBao’s live services and model post-training. Outstanding interns who meet the criteria may also have the opportunity to receive "DouBao shares."

Waiting until graduation to join might already be too late.

Each generation has its own opportunities. An annual salary of nearly ten million yuan is enough to transform an individual’s life. When invested in an AI company’s R&D, it’s a preemptive bet on the next breakthrough. The outcome is still undecided—but missing out on key talent could cost even more.

No one knows how long this wave of high salaries will last. When people talk about era-defining opportunities, they often think of entrepreneurship, fundraising, and IPOs. But for today’s young generation, the opportunity has become as concrete as a job offer.