01
Persistent agents
Maintaining context, state, and task continuity across long-running human–agent work.
CS PhD · University of Texas at Dallas
I build collaborative AI agents that unfold over time.
I study how agents persist, learn from experience, and spend intelligence where it matters.
Previously SMU · THUNLP · ModelBest · OpenBMB
Research threads
Researcher-builder · long-horizon agents, real-world evaluation, and human–AI collaboration.
01 / Current focus
What I am trying to understand now.
The common thread is continuity: how an agent keeps its bearings, learns from experience, and adjusts as the task, context, and relationship with a person evolve.
02
Learning from interaction
Turning real feedback, corrections, and outcomes into signals that can improve future behavior.
03
Adaptive agent systems
Helping agents adjust their behavior and effort as tasks, feedback, and collaboration with people evolve.
02 / Selected work
Things I have published.
01
DSDBench
EMNLP 2025 Main · Oral
Why Stop at One Error? Benchmarking LLMs as Data Science Code Debuggers for Multi-Hop and Multi-Bug Errors
A data-science debugging benchmark for tracing multi-hop logical errors and resolving multiple bugs in realistic code.
02
MatPlotAgent
ACL 2024 Findings
Method and Evaluation for LLM-Based Agentic Scientific Data Visualization
A visual-feedback agent that turns natural-language requests into scientific visualizations, then uses the rendered chart to repair its own code.
03
03 / Writing
A working notebook.
04 / Background
The path so far.
A few places where the work took shape.
University of Texas at Dallas
CS PhD
Beijing Language and Culture University
M.Eng. in Computer Science and Technology
Sichuan University
B.Eng. in Computer Science and Technology
05 / Contact