Jingyuan Liu
Hi! I'm Jingyuan Liu(刘静远), a sophomore student at Fudan University majoring in Software Engineering. I will be an exchange student at University of Toronto in Fall 2025. My research interests lie in learning-based robotic manipulation, embodied AI, and world models. I am currently working as a student intern at the Fudan Vision and Learning Lab under the supervision of Zuxuan Wu, where I focus on developing algorithms for robotic manipulation tasks. I am passionate about advancing the field of AI and robotics through innovative research and practical applications. Besides my academic pursuits, I am also a football⚽ lover.
Education
Fudan University, Shanghai, China
2023 -- Present
B.S. in Software Engineering, expected June 2027
GPA: 3.71/4.00
Transcript
Research Interests
Learning-based Robotic Manipulation, Embodied AI, World Model
Experience
Fudan Vision and Learning Lab, Shanghai, China
2025.02 -- Present
Student Intern (Advisor:
Zuxuan Wu)
- Used Mimicgen (an autonomous data generation pipeline) to train and evaluate the robustness of Diffusion Policy for robotic manipulation tasks.
- Extended 3D Diffusion Policy to multi-task scenarios.
- Currently developing algorithms that utilize optical flow for enhanced visual representation and perception in robotic manipulation.
Fudan Public Performance and Information Research Center, Shanghai, China
2024.06 -- 2025.02
Student Intern (Advisor:
Junyu Niu)
Participated in the development of LLM for taxation applications, focusing on database construction and data analysis.
- Data collection, annotation and feature extraction using LLMs.
- Audio data processing.
Class Projects
- Robot Navigation System
Developing a robotic navigation system in ROS, implementing Dijkstra and A* algorithm for global path planning and VFH algorithm for local path planning and obstacle avoidance.
- Mini Raytracer for Image Rendering
Built a high-performance ray tracing engine based on TinyRayTracer, support basic shading and acceleration structures.
- LLM Evaluation Platform
Designed and implemented a web-based platform for dataset management and LLM evaluation.
Skills
- Programming Languages: C, Python, C++, Java
- Tools: Linux, Git, CMake
- Deep Learning: Proficient in building, training, and optimizing models using PyTorch
- English Proficiency: TOEFL 101, Duolingguo 145; Skilled in researching, comprehending, and synthesizing information from English academic literature and technical documentation.
Honors and Awards
- Fudan University Outstanding Student Award (Second Prize) 2024.12
- Honorable Mention in Mathematical Contest in Modeling 2025