Zhejiang Unbounded Matrix Technology

Zhejiang Unbounded Matrix Technology, operating as Unbounded Matrix, is an artificial intelligence startup established in 2024. The company secured over ten million RMB in an angel funding round in late 2024 or early 2025, led by Leaguer Group (力合金融) and the Yangzhou Angel Science and Technology Venture Capital Fund.

The founding team is composed almost entirely of PhDs in AI from Tsinghua University, with one from Peking University. Founder Cui Sen brings a formidable academic background to the venture, having graduated from Tsinghua University's prestigious Qian Xuesen Class and studied under Professor Zheng Quanshui, an academician of the Chinese Academy of Sciences. His doctoral research was deeply focused on trustworthy AI and large models, a specialization that directly informs the company's core mission. The team, including co-founders Dr. Cao Zhong and Dr. Ma Jingheng, has been actively engaged in AI research since 2016, with published work in trustworthy AI, large models, intelligent agents, and affective computing.

Unbounded Matrix is focused on the burgeoning science and technology education market, which is projected to see significant growth in China. The company's business model centers on developing and deploying an "AI Tutor." This product is envisioned as a sophisticated educational tool that combines multi-modal technology, efficient inference, and emotional interaction. The goal is to provide students with personalized guidance and emotional support alongside skills-based instruction, moving beyond simple knowledge transfer to foster critical thinking and problem-solving abilities. The company has already developed a range of AI courses for different age groups, including an AI research base course (ages 6-12), an AI general education online course (ages 10-16), and AI-supported programming courses (ages 5-14).

The core offering is built upon a proprietary technology called "Abductive Retrieval-Augmented Generation." This technology integrates reverse reasoning with real-time retrieval to overcome the limitations of traditional models in handling timely information, reportedly achieving millisecond-level response speeds. This technical approach aims to provide a more dynamic and responsive learning experience than a general-purpose AI model could offer. While the primary focus is on educational products for students, Unbounded Matrix is also exploring applications for its guided learning system in professional settings, such as corporate seminars, to simulate real-world decision-making scenarios.

Keywords: AI Tutor, artificial intelligence in education, science and technology education, trustworthy AI, large language models, affective computing, multi-modal AI, personalized learning, edtech, intelligent agents, retrieval-augmented generation, Cui Sen, Tsinghua University alumni, Leaguer Group, AI curriculum, professional training simulation, guided learning systems, K-12 AI education, online learning platforms, educational technology