I am a fifth-year Ph.D. candidate in the Department of Computer Science at Rutgers University, advised by Prof. Dimitris N. Metaxas. My research focuses on large language models and generative AI, with an emphasis on retrieval-augmented generation, information retrieval and agentic AI. My work centers on how AI systems retrieve, represent, and reason over complex information, with the goal of improving their reliability, efficiency, and scalability.
Before joining Rutgers, I received my M.S. degree in Computer Science from Syracuse University in 2022 and my B.Eng. degree from Jilin University in 2020.
I am currently seeking full-time Research Scientist, Applied Scientist, or Machine Learning Engineering opportunities starting in 2027.
Education
Ph.D. Candidate in Computer Science
M.S. in Computer Science
B.Eng. in Logistics Engineering
Experience
Research Assistant
- Advisor: Dimitris N. Metaxas
- Agentic GraphRAG for noisy and incomplete KGs. [ICML 2026]
- Confidence-Guided Multimodal RAG. [EMNLP 2026 Findings]
- Pooling-Induced Collapse & Semantic Shift in Long-Text Retrieval. [arXiv 2026]
- Game-Theoretic Confidence Elicitation for LLM Metacognition.
Research Intern
- Works with Yangmin Ding & Shaobo Han.
- Efficient GraphRAG for Domain-Specific QA. [IJCNLP-AACL 2026]
Research Assistant
- Advisor: Sucheta Soundarajan
- Fairness in Social Network.
Selected Publications
Betting on Truth: Shaping LLM Metacognitive Confidence through Game-Theoretic Incentives and Batch Contrast.
In arXiv,
2026.
Pooling and Semantic Shift: the Fundamental Challenge in Long Text Embedding and Retrieval.
In arXiv,
2026.
CLIMB: Confidence-Guided Complementary Evidence for Multimodal Retrieval-Augmented Generation.
In EMNLP Findings,
2026.
BG-RAG: Concept-Mediated Bipartite Graphs for Scalable RAG.
In IJCNLP-AACL,
2026.
INMS: Memory Sharing for Large Language Model based Agents.
In IJCNLP-AACL,
2026.
Task-Aligned Tool Recommendation for Large Language Models.
In IJCNLP-AACL,
2025.
A-MEM: Agentic Memory for LLM Agents.
In NeurIPS,
2025.
Vector Retrieval with Similarity and Diversity: How Hard Is It?.
In arXiv,
2024.
2k-Vertex Kernels for Cluster Deletion and Strong Triadic Closure.
In Journal of Computer Science and Technology,
2023.
Contact
- hanggao [dot] gh [at] gmail [dot] com
- 617 Bowser Rd, Piscataway, NJ 08854