I am a Researcher at the Shanghai Academy of AI for Science (SAIS) , working with Prof. Siyu Zhu . My current research focuses on AI agents. I am interested in moving beyond content generation toward systems that can carry out useful tasks in collaboration with people. My background spans generative modeling, efficient inference, and computational imaging.
My long-term ambition is to contribute to artificial general intelligence that expands human capabilities. I hope to help build AI that reduces repetitive labor, broadens access to knowledge and tools, and enables new forms of collaboration and social organization. I am motivated not only by what AI can automate, but also by how its benefits can be broadly shared—giving people greater freedom to pursue creative, meaningful, and self-directed work.
I received my Ph.D. in Optical Engineering from Zhejiang University in June 2025, where I was affiliated with the State Key Laboratory of Modern Optical Instrumentation . My doctoral research focused on computational photography and image restoration. I also participated in the Pujiang National Laboratory Joint Training Ph.D. Program , with joint supervision from Prof. Yueting Chen and Prof. Tianfan Xue .
Before joining SAIS, I worked as a Senior Vision Algorithm Engineer at HONOR on diffusion-based mobile imaging. My experience in imaging research and industry informs my interest in building AI systems that work under real-world quality and efficiency constraints.
I am working on AI agents, with the goal of turning model capabilities into useful assistance with real-world tasks. I am interested in how these systems can complement human expertise and expand what people can accomplish.
My work involves language, image, and video generation, model evaluation, and efficient inference. I am interested in the trade-offs among output quality, reliability, and computational cost, particularly in practical generative systems.
My earlier research combines imaging models with learning-based reconstruction to address motion blur, low light, platform jitter, and atmospheric turbulence. Applications include mobile photography and high-resolution remote sensing.
The publications below reflect my work in computational imaging and restoration; AI agents are my current research focus.
IEEE Transactions on Geoscience and Remote Sensing, 2025
Physics-guided restoration for push-broom imaging using an equivalent jitter model.
SIGGRAPH Asia, 2024
A sim-to-real framework for event-based frame interpolation in low-light video.
AAAI, 2023
Modeling and restoring spatially varying local motion blur in real-world dynamic scenes.
Optics Express, 2023
Alternating spatial- and frequency-domain learning for atmospheric turbulence compensation.
Contributed to a cloud-based diffusion enhancement pipeline for telephoto super-resolution on HONOR Magic series smartphones. My work included model adaptation, data and image-quality evaluation, artifact suppression, and efficient inference.
Worked on RAW super-resolution and AI cloud enhancement during an internship, focusing on image restoration, model evaluation, and imaging pipeline optimization.
I welcome research discussions and collaborations on AI agents, generative models, and efficient AI systems—particularly work that connects technical progress with useful real-world applications.
zhangziran@sais.org.cn
Shanghai, China
Outside research, I enjoy playing basketball, learning about history, international relations, and economics, and exchanging ideas with people from different backgrounds.