Tech Talent Market Shift: AI Experience Paradox, Junior Hiring Down, Embodied Intelligence Jobs Up 57%

In the early stage of an industry, hiring demands often present paradoxical phenomena. Some companies require AI job candidates to have more than five years of experience using AI products, but ChatGPT was only released at the end of 2022, less than four years ago. Other companies recruit senior HR managers with an age limit of 30 years old, requiring a background from top overseas universities while also demanding a deep understanding of human nature. The global sales head position is required to be no older than 45, yet some founders making this demand are themselves 46 years old. Some entrepreneurs hope to find a 'Joseph Tsai of 20 years ago,' but such a talent profile hardly exists in reality. Data shows that since ChatGPT officially launched at the end of 2022, the share of senior positions (more than five years of experience) in white-collar recruitment has increased slightly, mid-level positions (one to five years) have risen modestly, while junior positions (zero to one year) have continued to decline. By September 2025, employment among software developers aged 22 to 25 had fallen nearly 20% compared to the peak at the end of 2022. Companies are increasingly paying for uncertainty, prioritizing the recruitment of talent capable of judging direction, decomposing boundaries, and proactively intercepting risks, while entry-level positions are shrinking as basic work is replaced by AI. Job boundaries are being shattered. Alibaba Cloud has created a position for 'AI product design front-end engineer,' requiring candidates to concurrently handle product, design, and front-end development. Some back-end R&D staff at Cainiao International have been forced to transition to full-stack after training. Alibaba Group has formed a number of full-stack teams, pushing front-end, back-end, and testing roles toward full-stack engineers. In the embodied intelligence field, the most sought-after are those with three to five years of vertical experience and hands-on exposure to actual robots. Those with one to two years of experience also have opportunities, but senior talent with more than ten years of experience in internet or autonomous driving fields who have not entered the field early will find it difficult to obtain core management positions at leading companies by 2026. In 2026, embodied intelligence data positions suddenly saw price increases. Previously, few companies hired data platform or data engineers on a large scale, but model iteration has increased demand for data, and well-funded enterprises have begun self-developing algorithm models, driving up the value of such talent. In the AI hardware direction, Huaqiangbei has already enabled AI glasses to be selected from solution provider PPTs, design finalized, and bill of materials calculated, allowing brand owners to decide production quantities and achieve rapid delivery. The cost of hardware production is extremely low, and those adept at identifying user needs and mastering traffic have become the fastest-rising group in the market. The labeling of talent value in the market is shifting. Dreame Technology has been regarded in recent years as a 'Whampoa Military Academy' in the robotics field, with several senior executives starting their own ventures and securing large early-stage financing. Former Dreame China CEO Guo Renjie founded Lexiang Technology, which completed three angel rounds totaling about RMB 500 million; former Dreame humanoid robot business head Yu Chao founded Luming Robot, completing nearly RMB 200 million in financing within six months. DJI's ecosystem has also spun off talent: former consumer drone division head Tao Ye and others founded Bambu Lab, with a Series A round of RMB 30 million; former battery business head Wang Lei founded EcoFlow, with a Pre-A round of RMB 30 million. The income gap continues to widen. In the 2026 cohort of top AI campus recruitment, outstanding students can receive annual salaries of RMB 1 million to 2 million, with top talent reaching RMB 3.5 million to 3.8 million, and daily internship compensation for core positions ranging from RMB 4,500 to 4,800. Meanwhile, roles such as junior engineers and older P6/P7 employees in peripheral business units at large tech firms are being compressed by AI. The value of projects is beginning to outweigh individual resumes; in the embodied intelligence industry, candidates who have participated in publicly recognized and substantial products attract attention across the entire sector. Talent flows follow three paths. First, a surge toward AI hard-tech hubs. In the first half of 2026, total financing in the domestic embodied intelligence track reached RMB 93.5 billion, a fivefold increase from the first half of 2025, with 322 financing deals, up 137% year-on-year. During the spring recruitment period, demand for robot algorithm engineers increased by 57% year-on-year, and recruitment for embodied intelligence-related positions surged 15-fold, with average monthly salaries close to RMB 62,000. Over the past two years, nearly 40 core executives from the autonomous driving sector have crossed over into embodied intelligence, including former Li Auto vice president of autonomous driving R&D Lang Xianpeng founding Kunlun Xing Robot, former XPeng chief scientist Guo Yandong founding Zhi Pingfang, and former Horizon VP Yu Yinan founding Vita Dynamics. At least 15 former executives from Huawei have left to start or join embodied intelligence companies. Second, talent flows to places with the shortest industrial loops. Beijing focuses on the 'brain' and 'cerebellum' of embodied intelligence, Shanghai can assemble core components of humanoid robots within a 150-kilometer radius, and Shenzhen offers supply chains, solution providers, and mass production experience. Embodied intelligence companies headquartered in Beijing or Shanghai set up branches in Shenzhen, and overseas teams also choose to build hardware teams in China. Third, talent moves along labels that enhance resumes. Talent flows from BAT to DJI, Insta360, Dreame, and leading embodied intelligence companies, shifting from single-point roles to composite positions such as agent, go-to-market, real machine, full-stack, and commercialization.