.png){kind=logo}
Yue Zhao
Junior Machine Learning Researcher specializing in multimodal LLMs and computer vision
Connect with Yue
Yue already has a relationship with Reval, so a warm intro from us gets a much better response than cold outreach.
Recommended
Already have an account?
About
LLM/multimodal systems builder who developed DuetGen, a practical multimodal interleaved text-image generation system using a decoupled MLLM planner and video-pretrained diffusion transformer for high-quality image generation with step-wise alignment. Built a 298K-sample interleaved dataset across 8 domains/151 subtasks and deployed a GPT-5-based automated evaluation framework; also has LangChain-based multimodal agent orchestration experience with custom state management and reliability testing.
Experience
Education
Key Strengths
- Built and deployed a production-oriented multimodal interleaved generation system (DuetGen)
- Architected decoupled MLLM planning + diffusion image generation with step-wise alignment for controllability
- Designed two-stage training to prevent error amplification and improve long-horizon stability in multi-step generation
- Created large-scale interleaved multimodal dataset (298K samples; 8 domains; 151 subtasks) plus two test sets
- Implemented GPT-5-based automated evaluation covering multiple quality/consistency dimensions
- Strong workflow/agent reliability practices: structured I/O, module-level tests, branch/edge-case coverage, system-level metrics
- Effective cross-functional communication translating evaluation metrics and experiments for finance/operations stakeholders
Discover more candidates like Yue
Search across thousands of pre-screened, high-quality, high-intent candidates on Reval.
Search Talent