The post-95s from Deepseek, poached by industry leaders with millions in annual salary, step into the spotlight of the large model stage
TechFlow Selected TechFlow Selected
The post-95s from Deepseek, poached by industry leaders with millions in annual salary, step into the spotlight of the large model stage
They were in tune with their era.
Navigating Web3 tides with focused insightsAuthor: Rao Fuying,TechFlow
Image source: Generated by Wujie AI
Behind Deepseek's explosive growth throughout the Spring Festival lies a group of "enigmatic geniuses" (as described by Jack Clark, former policy lead at OpenAI).
Liáng Wénfēng, founder of DeepSeek, once provided a rough profile of these employees in an interview with 36Kr: "They are mostly recent graduates from top universities, fourth- or fifth-year PhD students who interned before graduation, and some young professionals who graduated just a few years ago. In today’s AI industry, elite education and high academic qualifications have become standard."
The "2024 Report on the Profile of Chinese Artificial Intelligence Entrepreneurs" shows that AI entrepreneurs are highly educated and young, with over 70% holding master’s degrees or higher. They come from more than 400 universities worldwide, led by Tsinghua University and Peking University.
Moreover, post-95s now stand center stage. According to the "Global AI Talent Tracking Report" released in 2024 by Macro Polo, a U.S.-based think tank, among more than 30 prominent Chinese-founded AI startups in the United States, half of the founders are post-95s, including prodigy Guo Wenjing and Alexander Wang, whose net worth exceeds $1 billion.
In China, key developers behind DeepSeek-V2 such as Luo Fuli, Zeng Guoyang, CTO of MiniMax Intelligent, and Jiang Yuchen, founder of Waveform Intelligence, are also all post-95s.
"There are approximately 200 million working-age post-95s in China (aged 17–29). This generation grew up during a period of rapid economic growth, enjoying affluent lives and tending to value work-life balance," notes the report "China City Talent Attraction Rankings: 2024."
Therefore, it is no coincidence that post-95s have become central figures in the wave of large models—this stems closely from both historical trends and personal choices.
Genius Teens in the Spotlight
Data provided by iMedia Research (iiMedia Consulting) shows that China's AI large model market reached about 29.416 billion yuan in 2024 and is expected to surpass 70 billion yuan by 2026, demonstrating strong momentum.
The 2024 Government Work Report proposed launching an "Artificial Intelligence+" initiative, emphasizing AI's importance in the digital economy and accelerating high-level self-reliance and strength in science and technology. Large models and AI are currently star sectors, and those entering the field in recent years have done so at just the right time.
As widely known, AI is a high-barrier industry requiring absolute "hard skills." Looking at young talents stepping into the spotlight of the large model arena in recent years, most boast impressive academic backgrounds and strong technical capabilities.
Luo Fuli, rumored to have been poached by Xiaomi with a multi-million-yuan annual salary, is a rising star in the large model community. She completed her undergraduate studies in computer science at Beijing Normal University and was then admitted to the Institute of Computer Language at Peking University for graduate studies without taking entrance exams.
After earning her master’s degree, Luo Fuli joined Alibaba DAMO Academy's Machine Intelligence Lab through the "Alibaba Star Program," where she took charge of the AliceMind open-source project and led the development of the multilingual pre-training model VECO. The "Alibaba Star Program" represents Alibaba’s highest level of campus recruitment, with an average of fewer than 20 people hired annually and an acceptance rate below 1%. As Alibaba promotes, “Each cohort of Alibaba Stars consists of top-tier talent selected from thousands.”
As early as 2019, Luo Fuli made headlines after publishing eight top-journal papers within one year (two as first author), advancing developments in NLP technology through her proposed word sense disambiguation method and text style transfer framework.
In 2022, Luo Fuli left Alibaba to join幻方量化 (Huanfang Quant), focusing on deep learning strategy modeling and algorithm research. Later, she moved to DeepSeek as a deep learning researcher, contributing to the development of DeepSeek-V2.
Luo Fuli’s rise reflects both favorable timing and her exceptional professional competence. Another figure similarly hailed as a "prodigy girl" is Guo Wenjing, founder of Pika. She attended the competition-focused class at Hangzhou No. 2 High School. Her mother, Guo Xiaoli, is a distinguished alumna of MIT, while her father, Guo Huaqiang, is the actual controller of Xindaya Company.
In her first year of high school, Guo Wenjing won first prize in the Zhejiang provincial division of the National Olympiad in Informatics for Secondary School Students. She later participated in the U.S. National Olympiad in Informatics Open Competition for two consecutive years, winning first place both times.
In 2016, Guo Wenjing became the first student from Zhejiang Province to be admitted early to Harvard University’s undergraduate program. While pursuing her PhD at Stanford, Guo realized existing AI products were too complicated to use, prompting her to drop out and found Pika Labs, aiming to build an AI tool that simplifies video creation.
Venture capitalists in Silicon Valley quickly responded. Within a year, Pika secured Series B funding totaling $135 million, doubling its valuation to $470 million. Six months after founding, Guo launched Pika 1.0, enabling users to generate and edit videos in various styles—including 3D animation, anime, cartoons, and films—with zero barriers to entry.
The product went viral immediately upon release, gaining widespread attention across domestic and international tech circles. Alongside Pika’s success, her father’s company Xindaya surged—rising for six consecutive trading days over eight days, reaching a market capitalization of 8.3 billion yuan, a record high. The company even issued a statement clarifying it had not invested in Pika, but the story of a father benefiting from his daughter’s fame became widely celebrated.
Beyond Luo Fuli and Guo Wenjing, genius teens abound in the AI world. Zeng Guoyang taught himself programming at age 8, earned admission to Tsinghua University via an Olympiad championship, and was dubbed a "young prodigy" by the public.
During university, Zeng joined Tsinghua’s Natural Language Processing (NLP) Lab, becoming one of the earliest researchers in large models. He later joined the WenYuan Chinese pre-training model team as a core member. In 2020, he played a leading role in developing CPM-1, China’s first large-scale Chinese language model, sparking significant industry discussion.
After graduation, he co-founded MiniMax Intelligent with his advisor Liu Zhiyuan, serving as CTO at age 24. In May 2024, Zeng led his team in developing the "MiniCPM" model, praised by the industry as a "performance powerhouse."
They Are in Sync With Their Era
Beyond these already well-known post-95s, new faces from this generation continue to flood into AI-related entrepreneurial fields.
Jiang Yuchen, founder of Waveform Intelligence, is a post-95 PhD graduate in artificial intelligence from ETH Zurich. Less than a year after founding her company, she developed "Wawa Writing," an AI writing product for web novels powered by Weaver. Within a year of launch, the product attracted over 300,000 registered users, generated nearly 20 billion characters, and saw paying users spend an average of over six hours per day using it.
Song Yachen, born in 1998, focuses on 3D large models. A graduate of Johns Hopkins University, he previously worked at SenseTime and MiniMax. At just 25, he founded VAST, a 3D large model company. Within a year and a half, VAST emerged as the leading player in its field, capturing over 70% of the market share.
Photon Computing, focused on photonic computing chips, was co-founded by two post-95s. One co-founder, Xiong Yunjiang, holds a master’s degree in computational science from the University of Chicago; the other, Cheng Tangsheng, earned his doctorate from Oxford University. Their startup targets AI computing power chips, serving AI inference and training scenarios in the era of large models.
It’s not only in China—globally, post-95 academic elites are emerging as backbone forces. Scale AI, once a sensation in the AI world, raised seven rounds of financing and achieved a valuation in the tens of billions, becoming a mythic AI startup. Its leader, Alexander Wang, born in 1997, began coding at age 5, enrolled at MIT at 18, and dropped out during his freshman summer to found Scale AI.
Today, Scale AI has become a global leader in AI data services. Alexander Wang himself ranks among the world’s richest under 30 and is considered one of Silicon Valley’s youngest entrepreneurial giants. At just 27, his net worth exceeds $1 billion, with some calling him “the next Musk.”
The post-95 generation is described as confident, optimistic, creative, and daring—a generation unafraid to express themselves, challenge norms, emphasize individuality, and pursue personalized lifestyles. As “digital natives,” they were exposed early to information society, embrace new technologies, and possess sharp insight and creativity.
LinkedIn conducted a survey on career attitudes among thousands of professionals across different generations in first-tier and “new first-tier” cities. It found that post-95s show higher openness toward emerging industries compared to those born in the 80s and 90s, particularly showing strong interest in advanced tech sectors like big data, cloud computing, AI, and robotics.
An AI recruiter told *AI Light-Year* that based on hiring trends, enterprise demand concentrates around a few universities with relevant labs, where PhD students are especially sought after. The recruiter believes China currently has urgent demand for algorithm pre-training talent, with most companies recruiting directly from core labs—many of which exist only at China’s top-tier institutions.
Still, “academic credentials are just a foot-in-the-door; ability is the iron rice bowl.” In an inherently high-barrier industry, degrees and intellect may only be the first step. DeepSeek’s hiring philosophy aligns with this view. Founder Liang Wenfeng told 36Kr that DeepSeek’s core team primarily consists of fresh graduates and current PhD students from top universities. Beyond institutional background, DeepSeek places great emphasis on candidates’ competition achievements and innovation potential.
Both Luo Fuli and Zeng Guoyang have publicly stated they prefer challenging tasks and dislike dull, repetitive work—believing prolonged comfort zones hinder growth.
Zhang Peng, founder of Zhipu AI, told *AI Light-Year* that he values two qualities above all: learning ability and innovative thinking. “Starting point matters, but it only reflects past accumulation. Future growth depends on learning ability.” Zhang revealed that interviews assess how quickly and effectively candidates absorb information in both technical and business domains.
The future is here. On the AI stage, post-95s who have already stepped into the spotlight will create even more possibilities.
Join TechFlow official community to stay tuned
Telegram:https://t.me/TechFlowDaily
X (Twitter):https://x.com/TechFlowPost
X (Twitter) EN:https://x.com/BlockFlow_News
