GALAXEA Founder Gao Jiyang: Catfish Effect, Zeng Guofan, Waymo vs. Momenta, and Why Xu Huazhe Left

Gao Jiyang, founder of Star Chart Robotics, embodies a style of extreme efficiency, engineering decomposition, and pragmatism—the opposite of the techno-romanticism seen in figures like Liang Wenfeng or Yang Zhilin. In this interview, recorded at a time when co-founder Xu Huazhe was about to leave the company, Gao reflected on his journey and philosophy. He says the robotics industry has an extremely long chain; sometimes you have to bury your head in the dirt. There is no fairy dust, only relentless execution.

Gao grew up in Shijiazhuang, learning early that diligence and pattern recognition could compensate for a lack of innate talent. In sixth grade, he crammed for a placement exam and ranked third. In high school, he chose physics competition and outworked peers: while others did a problem set once, he did it two or four times, categorizing question types until he could map any problem to a solution framework. This earned him a spot at Tsinghua, where he saw true geniuses like Yang Zhilin and Han Yanjun, reinforcing his humility and his method of learning from setbacks by reading biographies and history.

A pivotal moment came when he failed to get into a top US PhD program. Seeking inspiration, he read about Zeng Guofan, a Qing dynasty official who transformed from an idealistic scholar into a pragmatic leader who mobilized resources to achieve real-world results. Gao realized that the key to building something was not intellectual purity but the ability to rally people and resources. Soon after, an internship at SenseTime under Tang Xiao’ou exposed him to deep learning: he trained his first neural network and was awed by its ability to automatically discover patterns from data. Cycling through Tsinghua Science Park, he thought, “This will replace human programming—I must work on this.” That moment defined his AI path.

At USC, he set a target of finishing his PhD in four years, requiring 4-5 top-tier conference papers. He classified CVPR/ICCV papers into three types—defining a new problem, improving performance, and reducing cost—and wrote accordingly. He collaborated with lab mates, co-authoring to increase output. He graduated in 3.5 years. After graduation, he evaluated industries: AI had to be the foundational variable. He rejected advertising and AI-as-a-service models, choosing autonomous driving as the first embodiment of physical-world AI. Later, he founded Star Chart to tackle embodied intelligence, sticking to the same principle: AI must be the core differentiator.

Gao contrasts himself with geniuses like Yang Zhilin. Lacking innate brilliance, he relies on systems: break down goals, summarize rules, and push through difficulties. His graying hair is a testament to the grind. In robotics, romanticism is a luxury; you must bury yourself in countless engineering details. Gao’s story is one of turning setbacks into fuel, using diligence and structured thinking to navigate a field that demands both vision and grit.