About

Research, systems, and selected work.

Education

2023.09 - 2027.06
Bachelor, Automation
Xi'an Jiaotong University
Supervisor: Qingyu Yang (Dean, School of Automation, Xi'an Jiaotong University).
Coursework and interests include automatic control, intelligent control, robotics, reinforcement learning, multi-agent systems, computer vision, deep learning, and foundation-model applications.
Recipient of the Yuejie Program Scholarship.
2025.07 - 2025.09
Summer Visiting Program
University of California, Berkeley
Summer Session visiting student.
Completed AI and scientific-engineering application coursework together with academic English (A+).
Developed the Bi-Talk simultaneous-interpretation agent with international teammates and received an outstanding project distinction.

I am Qinchuan Cheng, an undergraduate student in Automation at Xi’an Jiaotong University. My research spans multimodal foundation models, LLM agent systems, AI for Science, world models, and embodied intelligence.

What interests me is not only model performance itself, but how intelligent systems connect to real environments, physical constraints, social mechanisms, and end-to-end engineering systems. For me, AI is not an isolated algorithmic tool. It is a way to reorganize knowledge, action, and creative work.

Current Work

  • First author and core developer of Ego2World, a video-compiled symbolic simulator for embodied-agent evaluation.
  • Core developer of LAW MASTER / MAP-Law, a legal consultation agent system with retrieval control and evidence-coverage reasoning.
  • Research contributor to TransRheo and a broader polymer-material LLM project for AI-driven scientific modeling.
  • Research member on lightweight world modeling and planning for autonomous driving.

Internship

Peking University Seal
  • Intern, GoTim Project Group, Peking University (2026.4.25 - Present)
  • Personal work: GoTim robotics agent development and software debugging

Snapshot

  • Xi’an Jiaotong University, B.Eng. in Automation, 2023-2027
  • UC Berkeley Summer Session, Visiting Student, 2025
  • TOEFL 94

Research Focus

  1. Multimodal Foundation Models
    Multimodal reasoning, symbolic grounding, long-horizon interaction, and task execution in complex environments.
  2. LLM Agents
    Retrieval, planning, tool use, state maintenance, and interactive workflows for domain-specific systems.
  3. AI for Science
    Scientific modeling, material research, physics-informed learning, and agent support for discovery workflows.
  4. World Models and Embodied Intelligence
    Autonomous driving, belief-state planning, predictive modeling, and structured simulation environments.