About

I’m Yujing Ju (pronounced as: Yoo-jing Joo [tɕy^˥˥y^˧˥tɕiŋ^˥˩]). I am currently a Doctor of Philosophy student in AI/ML at the University of Florida, supervised by Prof. Mei Liu.

In 2025 summer, I received my Honors Bachelor of Engineering degree in Robotics with First-Class Award from Heriot-Watt University, where I was advised by Prof. Wei Pang.

My research is dedicated to advancing General AI, with a primary focus on Representation Learning and Embedding Optimization. This work involves designing novel architectures for a range of foundational models, including large language models, graph neural networks, and diffusion models. My methodology integrates relational patterns, graph topology, and causal inference into the model’s architectural design, equipping systems with the capacity to reason over complex interactions and latent structures within high-dimensional data. The efficacy of these approaches is validated on challenging real-world applications, utilizing large-scale, privacy-sensitive datasets from electronic health records and recommendation systems to benchmark their performance and robustness.

Highlights

• Directed Graph Modeling: Improved academic paper recommendation precision by capturing overlooked directional citation relationships.
• Biomimetic Robotics: Developed a hexapod robot capable of efficient locomotion and object grasping using Vision based detection and motion planning.
• Hypergraph Modeling: Enhanced group recommendation accuracy by using hypergraph probability matrix.
• Open Source Contributions: Collaborated on graph-based mathematical model integration as part of a Florida State University research internship.

My Projects

Here are a few of my key research contributions:

1. Academic Paper Recommendation System
   Developed a module that captures directed interaction effectively and apply it to academic recommendation system through directed edge modeling directed citation relationship.
2. Hexapod Robot for Object Capturing
   Designed a lightweight, bio-inspired hexapod robot with advanced motion planning and vision-based recognition.
3. Hypergraph Modeling for Group Recommendations
   Introduced hypergraph probability matrix that captures higher-order interactions in group-buying scenarios.
4. Open Source GNNIP Library
   Contributed to a open source project integrating graph structure based modeling algorithms and enchencing computational efficiency.

Skills

While my CV contains a full breakdown of skills, here’s a brief snapshot:

• Robotics Development: Proficient in ROS, NVIDIA Jetson platforms.
• Data Science Tools: Familiarity with Hadoop, Spark, Pandas, Numpy, Scipy and Matlab.
• Programming: Advanced in Java, C, C++, Python, Javascript, CSS, and HTML.

For a detailed skill list, visit my CV page.

Recognition

• Innovation and Entrepreneurship Scholarship, 2022
• Outstanding Youth Volunteer Award, 2022

Connect with Me

Feel free to connect or explore my work through the following platforms:

GitHub: juyujing LinkedIn: Yujing Ju YouTube: @dyodn