Chao Huang

I teach machines to see, hear, and reason as one.

omni-modal LLMsmultimodal reasoningaudio-visual generation

Research Scientist, Tencent Hunyuan

I build multimodal systems that reason across vision, language, and sound — teaching large models to perceive, infer, and create across vision, language, and audio. My recent focus is on faithful cross-modal reasoning in LLMs and few-shot generation with diffusion priors. I value work that is measurable and deployable.

My research centers on omni-modal reasoning and generation, with a preference for leveraging pretrained foundation models (LLMs, diffusion) over task-specific architectures. I focus on problems where quality, and real-world scalability all matter.

Ph.D. Computer Science, University of Rochester · B.Eng, Nanjing University