Zengyi Qin is the Co-Founder and AI Research Lead at MyShell, a decentralized platform that leverages blockchain technology to enable AI-native applications. He is also a Ph.D. candidate in Computer Science at Tsinghua University, where his research focuses on multimodal machine learning and generative artificial intelligence. ^{[1] [4] [6]}

Education

Zengyi Qin completed his Bachelor’s degree in Automation Science and Electrical Engineering at Beihang University in 2018. He then joined Tsinghua University’s Department of Computer Science and Technology as a Ph.D. student under the guidance of Professor Yuxin Peng. During his doctoral studies, he was a visiting student at the Massachusetts Institute of Technology (MIT), working with Professors Antonio Torralba and Phillip Isola on computer vision and representation learning. He also undertook a research internship at Microsoft Asia’s Visual Computing Group, where he contributed to large-scale vision-language models under Dr. Baining Guo. ^{[1] [4]}

Career

Zengyi Qin’s career began with a focus on embedded systems and real-time perception during an internship at Horizon Robotics. This early experience laid the groundwork for his later work in decentralized systems and blockchain technology.

He then joined Microsoft Asia’s Visual Computing Group as a research intern, where he worked on vision-language pretraining techniques. His contributions in this role advanced the development of multimodal AI systems, which integrate text, images, and other data types. This research served as a foundation for his later exploration of decentralized technologies, including blockchain, as a means to ensure transparency and community-driven innovation in AI development.

In 2023, Qin co-founded MyShell, a decentralized platform that combines AI and blockchain technology to enable creators to build AI-native applications. At MyShell, he oversees the development of open-source foundational models and tools designed to empower users while promoting decentralized governance. His work at MyShell reflects his commitment to creating accessible and transparent AI ecosystems.

Qin’s academic research focuses on unifying multimodal data into cohesive AI frameworks. His publications, presented at conferences such as CVPR, NeurIPS, and ICCV, address topics like cross-modal retrieval, generative model efficiency, and human-aligned AI training. He has contributed to open-source projects like LLaVA (Large Language-and-Vision Assistant) and MiniGPT-4, which integrate vision and language models for prototyping AI applications.

Zengyi Qin advocates for open-source AI and decentralized platforms, emphasizing the importance of transparency and community involvement in AI development. He has expressed support for initiatives that prioritize ethical considerations and user empowerment in the deployment of AI technologies.

Qin’s research has been cited over 1,000 times, according to Google Scholar. He serves as a reviewer for several AI conferences and journals, contributing to the advancement of multimodal systems research. His open-source projects on GitHub have been utilized by developers globally. ^{[1] [2] [3] [4] [5] [6]}

Interviews

Open-Source AI and Decentralized Innovation #01

In a September 2024 interview on the YouTube channel Naveen K, Zengyi Qin elaborated on his perspectives regarding open-source AI, decentralized systems, and the evolving role of artificial intelligence in society. Speaking as part of the Future of AI series, Qin framed open-source models as critical drivers of collaborative progress, enabling researchers and developers to build upon shared advancements rather than reinventing foundational work. He cited projects like Stable Diffusion for image generation and MyShell’s Open Voice for audio synthesis as examples of open-source technologies achieving commercial viability, while emphasizing their role in democratizing access to advanced tools.

Qin also addressed the concept of Artificial General Intelligence (AGI), describing it as a long-term aspiration to automate complex, repetitive tasks and enhance productivity across industries. However, he acknowledged current limitations, noting that most AI systems remain specialized and require integration with domain-specific expertise to maximize their utility. In discussing AI development challenges, he underscored the primacy of high-quality data over computational scale or architectural complexity, advocating for meticulous data curation to mitigate biases and inefficiencies.

Turning to advice for emerging researchers, Qin encouraged a focus on interdisciplinary applications, such as AI in education or environmental science, rather than competing directly with large corporations in foundational model development. He argued that niche innovations within specific fields could yield transformative impacts while sidestepping resource disparities.

Throughout the conversation, Qin reiterated his support for decentralized frameworks, particularly MyShell’s integration of blockchain technology to ensure transparency and community governance in AI ecosystems. He positioned such platforms as counterbalances to centralized control, fostering ethical innovation by prioritizing open collaboration and user agency.

The interview reinforced Qin’s alignment with open-source principles, framing them not only as technical strategies but as mechanisms to democratize AI’s societal benefits. His remarks highlighted a pragmatic optimism, balancing academic rigor with a commitment to accessible, equitable technological advancement. ^[7]