Research & interests
The questions behind the work.
Scheduling & learning
Making decisions
with dependencies.
How should dependent tasks be assigned and ordered when computation and data movement share limited resources? My research focuses on deep reinforcement learning for workflow scheduling.
Agents & applications
Turning model capabilities
into useful tools.
I explore how language models, tools, and backend services can work together in AI agent applications, with an interest in the engineering decisions that make these systems usable.
Across both areas, I care about explicit assumptions, useful implementations, and experiments that can be inspected and reproduced.
Selected project
A closer look at what I’m building.
Research software · Java
A discrete-event simulation environment for scientific workflow scheduling. It brings workflow models, experiment configuration, and replayable evidence into a single research workflow.
- Model
- Workflows, compute, and data
- Experiment
- Explicit setup and validation
- Inspect
- Reports, traces, and replay
A little about me
Based in Changsha, China.
I’m a software engineering master’s student at Changsha University of Science and Technology, jointly trained at the National University of Defense Technology. I expect to graduate in 2027.
I mainly work with Python and Java. I enjoy turning research questions into experiments and ideas into software that people can use.