Coding Agent Research Intern

Focus on cutting-edge algorithm research and development in AI Coding Agents, fully automated game generation, and game world models, exploring key issues such as code generation, intelligent debugging, gameplay, and level generation. As a Coding Agent Researcher (Intern), you will participate in one or more of the following tasks: Develop LLM-based game coding agents to enable automatic code generation, game logic development, error diagnosis, automatic debugging, and iterative optimization. Explore the use of AI to fully automate the generation of game demos, core gameplay, levels, and other procedural game content. Track cutting-edge directions such as Coding Agents, code large language models, and world models, and conduct algorithm experiments, paper reproduction, and technological innovation. Drive the practical application of research outcomes in game development and interactive entertainment scenarios, and produce high-quality technical reports, papers, or patents. Collaborate closely with senior researchers and game development teams to jointly explore AI-driven game development and virtual world construction technologies.

Currently enrolled in a Master’s, or PhD program in Computer Science, Artificial Intelligence, Software Engineering, Machine Learning, or a related field. Familiar with Python, with solid programming skills and a strong foundation in deep learning; understand the basic principles of large language models and intelligent agents. Capable of independently conducting experiments, debugging code, and optimizing models, and able to complete the full workflow from solution design to experimental validation. Strong interest in game development, procedural content generation, and the application of generative AI in interactive entertainment. Good communication, collaboration, and technical presentation skills. Preferred Qualifications Familiarity with Unity, Unreal Engine, or other game engines; experience in developing Coding Agents, code large language models, tool use, or multi-agent systems is preferred. Experience publishing relevant papers at top-tier conferences or journals such as CVPR, ICCV, ECCV, ICLR, ICML, and NeurIPS is preferred.