Personal Statement for GSoC 2025 – Demonstrating Integration of open-webui with OpenVINO Model Server

[ArborsQ]

Dear @adrianboguszewski @mlukasze,

I am Gaoyuan Qin, currently pursuing a Bachelor's degree in Software Engineering at Swinburne University of Technology, with a prior academic foundation at Shandong University of Science and Technology. With deep technical experience in artificial intelligence, especially in large language models (LLMs), computer vision, and multimodal systems, I am enthusiastic to apply for the project “Demonstrating integration of open-webui with OpenVINO Model Server.”

My academic and research background makes me a strong fit for this project. At Shanghai Jiao Tong University, under the guidance of Prof. Mengjiao Wu, I contributed to a project that involved deploying and testing large-scale speech models to support AI-driven psychiatric consultation. This work required understanding model inference, API integration, and user experience in interacting with generative models—closely aligning with the challenges of integrating OpenVINO Model Server into a frontend UI like open-webui.

In addition, I have experience building intelligent systems involving YOLOv5/v8 models, CNNs, and multimodal data. I authored two research papers: one on multimodal emotion recognition using 3D-CNNs, and another on YOLOv8-based defect detection, where I contributed to model optimization and deployment workflows. These projects have developed my practical skills in Python and model deployment pipelines.

Beyond research, I have hands-on experience with software engineering projects involving end-to-end system design. My current project at Swinburne involves using Hadoop and LSTM to process driving data, and I also developed a task planning Android app, both of which reflect my ability to build usable software systems with clear user interfaces—key to improving and analyzing open-webui’s usability in this project.

What excites me most about this GSoC opportunity is the chance to work at the intersection of LLMs, inference optimization, and user interface design. I’m particularly interested in exploring the integration of chat and image generation endpoints using standard APIs, analyzing performance and usability gaps, and contributing documentation that can help other developers adopt this integration approach.

I believe my blend of academic rigor, research in LLMs and multimodal AI, and hands-on system development makes me well-suited to this project. I am confident in my ability to collaborate with mentors Dariusz and Milosz, and I am committed to delivering a polished and practical integration of open-webui with OpenVINO Model Server.

Thank you very much for considering my application.Can you please help me find the mentors?

Sincerely,
Gaoyuan Qin