Medical AI company "Meiyin Intelligence" secures multi - million - yuan seed round financing by predicting disease risks with large models | Exclusive from 36Kr

Text by | Hu Xiangyun

Edited by | Hai Ruojing

36Kr learned that Hangzhou Meiyin Intelligent Technology Co., Ltd. (hereinafter referred to as "Meiyin Intelligent") has recently completed a seed - round financing worth tens of millions of yuan. This round of financing was led by institutions such as Joyuan Asia, with the direct - investment fund of Hangzhou West Lake Government, West Lake Science and Technology Innovation Investment, participating as a follow - on investor. The raised funds will be mainly used for the research and development of the disease risk large model, core products, and commercialization.

It is reported that Meiyin Intelligent was incubated in the Peking University Science Park and focuses on using artificial intelligence technology to empower disease risk prediction and health management. The company's core products are a major disease risk prediction platform and a disease risk management AI agent based on the company's self - developed disease risk large model, providing AI - driven insurance and disease risk management solutions for people at risk of serious and chronic diseases. At present, the company's self - developed DP - LLM large model can support multi - modal medical data, quantify an individual's future disease risk, covering hundreds of diseases and tens of thousands of risk factors.

When talking about why the company focuses on the disease risk prediction scenario, Guo Xiaoyu, the founder and CEO of Meiyin Intelligent, explained that from a technical perspective, compared with the previous generation of discriminative medical AI, today's generative large models are "sequence prediction models", which have the ability to predict a person's future health status and disease progression based on their historical health data sequence. "At the policy level, strengthening early diagnosis, early treatment, and early rehabilitation is the strategic theme of the 'Healthy China 2030 Plan Outline'; at the individual level, personalized treatment and precise health management are also the current development trends in the medical industry."

To put it simply, in the past, without high - quality data and precise risk assessment models as support, most insurance products set strict notification and exclusion clauses, making it difficult for many people with pre - existing conditions to be covered. "However, with the support of our model, we can make more detailed divisions of the population and insurance items, and provide different insurance values to different groups of people. For example, some people at early - stage disease risk may be included, and healthy people may get better insurance benefits, thus expanding the scale of the insured population."

To achieve this, Meiyin Intelligent added a lot of "ingenuity" when fine - tuning the model. First of all, in order to enable the model to analyze an individual's disease risk more accurately, the company conducted comparative studies based on a large number of medical literatures. For example, it quantified the correlation between a certain risk factor and a disease, such as "how much the corresponding risk value increases when the risk factor increases". Another example is to observe the disease conversion rate of a certain indicator within a certain period in the future based on urban epidemiological survey data.

On the other hand, Guo Xiaoyu also talked about the "hallucination" problem of medical large models. "The common method in the industry is to build a knowledge base and a knowledge graph, which is also what we are currently focusing on. By building a knowledge graph of the correlation between disease risk factors and diseases and then combining the knowledge graph with the fine - tuned model, we can reduce the occurrence of 'hallucinations'." At the same time, the company is also cooperating with government departments and medical institutions to verify and improve the accuracy of the model using real clinical data.

In addition, from a market perspective, based on his years of experience in the medical AI field, Guo Xiaoyu believes that "it is difficult to expand the market by simply targeting hospitals and doctors." For medical AI products to generate greater value, they still need to be directly combined with individuals.

At present, Meiyin Intelligent mainly chooses commercial insurance payers as an entry point and cooperates with government departments and large insurance companies to empower relevant health insurance products with its model. While promoting commercial transformation through the "B - end technical service fee, C - end subscription, and risk - sharing model", it can reach more C - end users. In the future, the company will also try to develop more health management products. "The future of medical AI products must be in the C - end," Guo Xiaoyu believes.

In terms of the team, Guo Xiaoyu is a master from the Department of Computer Science and Technology of Tsinghua University and a doctoral candidate in artificial intelligence at Peking University, studying under Professor Cui Bin, the deputy dean of the School of Computer Science at Peking University and an IEEE Fellow. Guo Xiaoyu has more than 10 years of R & D and commercialization experience in the medical AI field. He was an early member and vice - president of a Hong Kong - listed medical AI company, leading the growth of the health insurance technology sector from 0 to over 2 billion yuan in premium scale. Co - founder Gan Shusheng (CTO) was a technical executive at Baidu, 360, Horizon Robotics, and several medical technology companies, specializing in AI engineering architecture and commercialization. Chief AI Scientist Yang Tong is a researcher, doctoral supervisor, and Young Yangtze River Scholar at the School of Computer Science at Peking University. He has published 100 CCF Class A papers, led the R & D of several core technologies of large models, and led the team to jointly develop the medium - scale inference model Tiny - R1 - 32B - Preview with 360, which approaches the full - blooded performance of Deepseek - R1 with only 5% of the parameters. Chief Scientist Advisor Professor Cui Bin is a Yangtze River Scholar of the Ministry of Education, an IEEE Fellow, the deputy dean and professor of the School of Computer Science at Peking University, and the director of the Institute of Data Science and Engineering at Peking University.

Investors' Views:

Lin Haizhuo, the founding partner and chairman of Joyuan Asia, said that the Meiyin Intelligent team combines top - notch academic backgrounds with practical experience. Its disease prediction large model has significant advantages in lightweight deployment and multi - modal fusion, and is expected to reshape the health insurance and health management ecosystem.

Zhao Chenyan, the investment director of West Lake Science and Technology Innovation Investment, said that the company's founding members have complete work experience in delivering commercial insurance disease prediction models. Meiyin Intelligent combines real - world data and AI modeling capabilities, and is committed to the efficient iteration of health insurance commercial closed - loop products, which is highly consistent with the development plan of Hangzhou's innovative digital health industry.