About

About

I am Runze Li, a full-time Research Assistant now at MiLab at Westlake University, advised by Professor Donglin Wang. My research focuses on embodied intelligence and robot foundation models, especially Vision-Language-Action models, world models, and generalist robotic agents for instruction understanding, implicit planning, and action generation.

Previously, I worked remotely with Professor Jordan Boyd-Graber at the University of Maryland on natural-language-guided image geolocation, where I helped develop the NAVIG framework and built data collection and evaluation pipelines for image geolocation with vision-language models. I also conducted research at DisLab at Nanjing University with Professor Wenzhong Li, working on deep learning methods for video super-resolution.

I received dual Bachelor’s degrees in Computer Science and Financial Engineering from Nanjing University.

Goal: Advancing toward general-purpose robotic agents through embodied intelligence.

Focus: Building robust, continual, and generalizable robotic systems with Vision-Language-Action models and world models.

Email: lirunze [AT] westlake.edu.cn

Selected Publications

Projects and Experience

  • Smart Assistant App Development and Deployment, May 2024 - Oct 2024
    I built backend services with FastAPI, integrated Google Play billing and subscription logic, implemented JWT authentication, deployed AWS EC2 services with SSL, and migrated payment processing to Supabase. The app was published on Google Play as Replify.

  • 19th Citi Financial Innovation Application Competition, Jan 2024 - Mar 2024
    I worked on AI-based equity-structure analysis for listed companies, using Graph Convolutional Networks and hierarchical graph pooling methods for risk classification. Our team ranked in the national Top 20.

My latest CV is available here: Resume.

Trending Tags