Hi! I am Xuning He, an undergraduate student in AI at Nankai University. My research interests include LLM & Agents, Embodied AI, and decision planning that combines large language models with optimization solvers.
πEducation
-
Sep 2023 - Present, Undergraduate Student in AI, Nankai University
- GPA: 3.825/4.0; Rank: 1/53
- Major courses: Machine Learning, AI Technologies, Fundamentals of Machine Vision, Principles of Automatic Control, Data Structures, Modern Control Theory, etc.
πSelected Publications
- Safe Truck-Drone Delivery in Urban Air Corridors. Submitted to IEEE Transactions on Intelligent Transportation Systems (SCI Q1); third author.
- Ye, H., He, X., et al. (2026). Meta Context Engineering via Agentic Skill Evolution. ICML 2026 (accepted). Code
π¬Research Projects
-
May 2026 - now,
Semantic-Level Cooperative Diversion Decision-Making for Intelligent Connected Vehicles
- Institution: Department of Automation, Tsinghua University
- Studies cooperative decision-making for multiple connected vehicles with semantic driving intentions in urban expressway weaving areas; builds a hierarchical framework of upper-level semantic planning by LLMs and lower-level precise decisions by optimization solvers.
- Role: Project leader.
-
Jan 2026 - Apr 2026,
Generative LLM-Driven Differentiable Embodied Co-Design
- Institution: College of Artificial Intelligence, Tsinghua University
- Explores a Meta-Embodied-Co-Design approach that breaks the dual-loop pipeline, uses LLM Coding to explore design space, and optimizes design efficiency and quality through mathematical formulations.
- Role: Project leader and main lead for the laboratory hardware direction.
-
Oct 2025 - Jan 2026,
Meta Context Engineering via Agentic Skill Evolution
- Institution: School of Intelligence Science and Technology, Peking University
- Proposed Meta Context Engineering (MCE), which co-evolves context skills and context artifacts to replace the heuristic generation-reflection framework in Context Engineering.
- Outcome: ICML 2026; achieved 5.6-53.8% relative improvement over SOTA methods.
-
Jun 2025 - Sep 2025,
Reinforcement Learning for Multi-Truck Multi-Drone Collaborative Delivery
- Institution: Swarm Intelligence and Pattern Recognition Research Group, Nankai University
- Built a MILP model with time-window constraints and an end-to-end reinforcement learning model, using graph attention networks for global information and a scoring network for task decisions.
- Outcome: Submitted to T-ITS (SCI Q1); third author.
-
Feb 2025 - Jun 2025,
Intelligent Bionic Seahorse Robot
- Institution: Swarm Intelligence and Pattern Recognition Research Group, Nankai University
- Led the tail grasping module and mechanical system; designed a multi-joint cable-driven underactuated grasping mechanism, completed 3D printing and hardware setup, and used CPG-based coupled control for autonomous navigation and target grasping with visual and infrared feedback.
- Outcome: National runner-up in the International Advanced Robotics and Simulation Technology Competition; Tianjin Undergraduate Innovation and Entrepreneurship Project.
-
Feb 2025 - May 2025,
Multi-Servo Home Service Robot
- Institution: Intelligent Perception and Human-Computer Interaction Laboratory, Nankai University
- Built a ROS-based workflow integrating voice interaction, navigation and obstacle avoidance, YOLOv11 object recognition, and robotic arm grasping.
- Outcome: Third Prize, China Robot Competition and RoboCup China Open.
πHonors and Awards
- National Scholarship
- Honor AI Person of the Year
- Nankai University Merit Student
- Excellent Teaching Assistant for Probability Theory and Mathematical Statistics, Nankai University
- Third Prize, China Robot Competition and RoboCup China Open
- National Runner-up, International Advanced Robotics and Simulation Technology Competition
- First Prize, North China Five-Province Robot Competition
- Meritorious Winner, Mathematical Contest in Modeling
πLanguage Proficiency
- Mandarin: Native
- English: Fluent, CET-6 505