Using a large model to create an AI technology tutor, "Unbounded Matrix" has completed an angel round of financing of over 10 million yuan. | 36Kr exclusive.
Written by | Wang Fangyu
Edited by | Su Jianxun
36Kr has learned that Yangzhou Borderless Matrix Technology Co., Ltd. (hereinafter referred to as "Borderless Matrix") recently completed an angel round of financing of over ten million yuan with a valuation of hundreds of millions of yuan. This round of financing is led by Lihe Finance.
Borderless Matrix is an artificial intelligence startup founded in 2024. The founder, Cui Sen, completed his undergraduate studies in Tsinghua Qian Xuesen Class, under the tutelage of Professor Zheng Quanshui, an academician of the Chinese Academy of Sciences. During his doctoral studies at Tsinghua, he has been deeply engaged in the field of trustworthy AI and large models.
The core goal of the company is to create an "AI Tutor" that integrates multimodal technology, efficient reasoning, trustworthiness, and emotional interaction. It not only teaches students technical skills but also provides personalized emotional companionship.
"AI technology will be the core key for every organization and individual to meet future challenges. Different from the traditional learning of mathematics, physics, and chemistry, we want to turn AI into a basic thinking tool that everyone can master," Dr. Cui Sen told 36Kr.
At present, the first product of Borderless Matrix is under development and is expected to be officially launched in the middle of 2025.
Cui Sen said that the application scenarios of AI large models often involve not just one technology, but a comprehensive project of a series of technologies. The AI Tutor large model of Borderless Matrix requires breakthroughs in various links such as knowledge retrieval, reasoning optimization, and explanation verification, which is difficult for a single general AI large model to achieve.
For example, abductive retrieval-enhanced generation technology. It combines abductive reasoning with a real-time retrieval mechanism to improve retrieval efficiency and accuracy, breaking through the limitations of traditional retrieval-enhanced generation in terms of information timeliness and achieving a response speed at the millisecond level.
In contrast, traditional retrieval-enhanced generation often struggles in complex environments due to a lack of cognitive autonomy, making it difficult to flexibly respond to real-time needs, especially with obvious deficiencies in autonomously determining "when to retrieve" and "what to retrieve".
Borderless Matrix has just been registered and established, but the team members have been closely following the cutting-edge of AI innovation since 2016, including multimodal, meta-learning, and efficient reasoning of large models, with rich industry experience.
The core founding team of the company are all PhDs in the AI direction from Tsinghua University. They have published many academic papers in well-known domestic and international conference journals and have years of accumulation in the fields of trustworthy AI, large models, agents, and emotional computing. Among the co-founders are Dr. Cao Zhong, Ma Jingheng, Dr. Hua Xiayang, Dr. Hong, who are also from the AI direction of Tsinghua, and Dr. Meng Delan from Peking University.
In the present era where technology has become the primary productive force, the simple exam-oriented education can no longer meet the needs of talent cultivation, and the status of science and innovation education is on the rise. Institutions such as iResearch and Duojing Education predict that the domestic science and innovation education market will grow at a rate of nearly 20% in the next five years, with a long-term scale of nearly 100 billion yuan.
"The application boundary of technology is usually not fixed but is continuously evolving and expanding," Cui Sen said. The team is currently building science and technology education products based on the "AI Tutor" technology, and at the same time, it is also trying to apply the guided learning system to seminars to simulate the decision-making process in actual work scenarios and improve the problem-solving ability and critical thinking of professionals.
In the next step, Borderless Matrix will inject more resources into technology research and development, selectively follow the cutting-edge large model technologies and focus on breakthroughs, accelerate the improvement and optimization of the functions of the AI Tutor product, and also explore more application scenarios of the AI Tutor and build a diversified application ecosystem with more partners.
Investor's Perspective:
Shen Kang, the investment director of Lihe Finance, said that in the foreseeable future, AI applications will undoubtedly bring breakthrough efficiency improvements to many industries and even bring disruptive changes in production relations. In the future, the "AI +" concept will inevitably give rise to many new business forms. We are optimistic about Dr. Cui's team's technical strength and entrepreneurial spirit. The core technologies such as trustworthy AI that the team has accumulated and mastered in the past are precisely the key capabilities that are urgently needed for the empowerment of various industries by AI agents at present. Dr. Cui's team chose to enter from AI + education, and quickly launched products and continuously iterated and updated, and the business model has also been effectively verified.