CV

Contact

• Jiaxing Zhang
• Research Scientist
• Bellevue, WA
• Email: tabzhangjx@gmail.com
• Phone: +1 (551) 290-9454
• Website: tabzhangjx.github.io
• GitHub: github.com/jz48
• LinkedIn: linkedin.com/in/jiaxing-zhang-45593b156

Education

• PhD in Informatics, New Jersey Institute of Technology, Newark, NJ, US (Sep 2020 - May 2025)
• Bachelor of Computer Science (Honored Science Program), Xi’an Jiaotong University, Shaanxi, China (Sep 2016 - Jul 2020)
• Exchange Student, University of Minnesota Twin Cities, MN, US (Jan 2019 - Jul 2019)
• Summer School, University of Alberta, Edmonton, Alberta, Canada (Jul 2017 - Aug 2017)

Experience

• Research Scientist, TikTok Inc., Bellevue, WA (Jun 2025 - Present)
  • Built end-to-end multimodal LLM (LVLM) systems for video moderation.
  • Optimized training/inference stack with HF Accelerate, torchrun, DeepSpeed, Megatron-LM, FSDP/ZeRO, mixed precision, and FlashAttention-2.
  • Extended model architectures across Qwen2.5-VL, InternVL, LLaMA, and DeepSeek-R1 (including MoE routing, vision projector modules, and 3D-RoPE).
  • Drove multi-stage pipelines across pre-training, fine-tuning, SFT, and post-training.
  • Collaborated with moderation policy/product/infra teams to ship reliable serving.
  • Selected works:
    • Revisiting Attention Sink in Large Vision-Language Models: Insights from Video Moderation (under review)
    • Teaching LVLMs to Rank via Capability-Shift Tuning (under review)
    • RB-FT: Rationale-Bootstrapped Fine-Tuning for Video Classification (under review)
• Senior Data Scientist (Full-time), Walmart Global Tech, Hoboken, NJ (Sep 2024 - May 2025)
  • Built recommendation systems using regularization/penalization, LLM-augmented ranking, and KG features.
  • Designed business-KG-driven RAG + agentic framework for enterprise reasoning and actions.
  • Selected work:
    • [SIGKDD 2025] Is Your Explanation Reliable: Confidence-Aware Explanation on Graph Neural Networks.
• Applied Scientist Intern, Amazon, Bellevue, WA (Jun 2024 - Aug 2024)
  • Developed a Smart Model Re-Training framework to handle time-based data distribution drift.
• Research Assistant (Explainable AI on GNN), LabDaRL, New Jersey Institute of Technology, Newark, NJ (Sep 2022 - May 2025)
  • [NeurIPS 2024] RegExplainer: Generating Explanations for Graph Neural Networks in Regression Task.
  • [ICML 2024] Interpreting Graph Neural Networks with In-Distributed Proxies.
  • [SIGKDD 2023] MixupExplainer: Generalizing Explanations for Graph Neural Networks with Data Augmentation.
• Research Assistant (Natural Language Processing), SPACE Lab, New Jersey Institute of Technology, Newark, NJ (Sep 2020 - Jul 2022)
  • [ESEC/FSE 2023] Commit-level, Neural Vulnerability Detection and Assessment.
  • [ESEC/FSE 2023] DeMinify: Neural Variable Name Recovery and Type Inference.
• Research Assistant Intern (Computer Vision), Park Lab, University of Minnesota Twin Cities, Minneapolis, MN (May 2019 - Aug 2019)
  • Worked on 3D reconstruction using projective geometry and CNNs for point cloud generation.
  • Developed image segmentation and object detection methods to improve reconstruction quality.

Publications

Is Your Explanation Reliable: Confidence-Aware Explanation on Graph Neural Networks

Jiaxing Zhang, Xiaoou Liu, Dongsheng Luo, Hua Wei. 2025. Is Your Explanation Reliable: Confidence-Aware Explanation on Graph Neural Networks. Proceedings of the 31st ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2025).

GE-Chat: A Graph Enhanced RAG Framework for Evidential Response Generation of LLMs

Longchao Da, Parth Mitesh Shah, Kuan-Ru Liou, Jiaxing Zhang, Hua Wei. 2025. GE-Chat: A Graph Enhanced RAG Framework for Evidential Response Generation of LLMs. IJCAI 2025 Demo Track (arXiv:2505.10143).

RISE: Radius of Influence based Subgraph Extraction for 3D Molecular Graph Explanation

Jingxiang Qu, Wenhan Gao, Jiaxing Zhang, Xufeng Liu, Hua Wei, Haibin Ling, Yi Liu. 2025. RISE: Radius of Influence based Subgraph Extraction for 3D Molecular Graph Explanation. International Conference on Machine Learning (ICML 2025).

RegExplainer: Generating Explanations for Graph Neural Networks in Regression Tasks

Jiaxing Zhang, Zhuomin Chen, Hao Mei, Longchao Da, Dongsheng Luo, Hua Wei. 2024. RegExplainer: Generating Explanations for Graph Neural Networks in Regression Tasks. Advances in Neural Information Processing Systems 37 (NeurIPS 2024), 79282-79306.

LLMExplainer: Large Language Model based Bayesian Inference for Graph Explanation Generation

Jiaxing Zhang, Jiayi Liu, Dongsheng Luo, Jennifer Neville, Hua Wei. 2025. LLMExplainer: Large Language Model based Bayesian Inference for Graph Explanation Generation. AIAgent4IR 2025 Workshop (in conjunction with KDD 2025).

Interpreting Graph Neural Networks with In-Distributed Proxies

Zhuomin Chen, Jiaxing Zhang, Jingchao Ni, Xiaoting Li, Yuchen Bian, Md Mezbahul Islam, Ananda Mohan Mondal, Hua Wei, Dongsheng Luo. 2024. Interpreting Graph Neural Networks with In-Distributed Proxies. International Conference on Machine Learning (ICML 2024).

MixupExplainer: Generalizing Explanations for Graph Neural Networks with Data Augmentation

Jiaxing Zhang, Dongsheng Luo, Hua Wei. 2023. MixupExplainer: Generalizing Explanations for Graph Neural Networks with Data Augmentation. Proceedings of the 29th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD 2023).

DeMinify: Neural Variable Name Recovery and Type Inference

Yi Li, Aashish Yadavally, Jiaxing Zhang, Shaohua Wang, Tien N. Nguyen. 2023. DeMinify: Neural Variable Name Recovery and Type Inference. Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE 2023).

Commit-level, Neural Vulnerability Detection and Assessment

Yi Li, Aashish Yadavally, Jiaxing Zhang, Shaohua Wang, Tien N. Nguyen. 2023. Commit-level, Neural Vulnerability Detection and Assessment. Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE 2023).

Talks

RegExplainer: Generating Explanations for Graph Neural Networks in Regression Task

March 01, 2024

Poster at Learning on Graphs Conference, New York, NY, USA

Address Out-Of-Distribution Problem in Explainable AI in Graph Neural Networks

August 09, 2023

Talk at 29TH ACM SIGKDD CONFERENCE ON KNOWLEDGE DISCOVERY AND DATA MINING, Long Beach, California

Teaching

IS218: Web Application Development

Services and Membership

• Reviewer/PC: NeurIPS, ICML, ICLR, KDD, AISTATS, WACV, PKDD, AAAI (PC)
• President, Chinese Students and Scholars Association (CSSA) at NJIT (2023 - 2025)
• VP of Student Involvement Committee, 32nd CAST-GNY Annual Conference (Oct 2024)
• Professional Member, ACM (2025 - 2026, invited by ACM President Yannis Ioannidis)

Skills

• Tools and Languages: Python, C/C++, Java, C#, JavaScript, Matlab, Anaconda, Docker, Linux, Git, LaTeX, Markdown, S3, Merlin
• Framework and Infra: PyTorch, TensorFlow, DeepSpeed, Megatron-LM, FSDP/ZeRO, HF Accelerate, torchrun, CUDA, Ray
• Models and Methods: LVLMs (Qwen2.5-VL, InternVL, LLaMA, DeepSeek), MoE, 3D-RoPE, vision projector, retrieval RAG, KG (GraphSAGE/TransE), temporal modeling, calibration, post-training (RLHF/DPO)
• Machine Learning: NLP, code analysis, GNN, explainable AI, information theory, time series, spatio-temporal graph, traffic flow prediction, LLM
• Communication: English, Chinese