About me

I’m Jiaxing Zhang, a Research Scientist at TikTok in Bellevue, WA.

I received my PhD in Informatics from New Jersey Institute of Technology in May 2025, where I was advised by Prof. Hua Wei and Prof. Michael Lee. I obtained my bachelor’s degree in Computer Science (Honored Science Program) from Xi’an Jiaotong University in 2020.

Research

My research focuses on two connected directions: trustworthy machine intelligence for structured data, including graph neural network explainability, robustness under distribution shift, and evidence-grounded graph RAG; and efficient multimodal foundation models, including LVLM training and post-training, long-video understanding, information-aware frame sampling, mixture-of-experts routing, and scalable ML systems. Across both directions, I aim to build models that preserve decision-critical information while remaining reliable, interpretable, and efficient in real-world applications.

I actively welcome research collaborations in trustworthy graph learning, GNN/LLM/VLM explainability, multimodal and video understanding, model architecture and training optimization, and scalable ML systems. I’m also open to working with motivated, well-prepared interns.