Yun-shiuan Chuang is a Staff AI Research Scientist with nine years of experience bridging cognitive science and cutting-edge LLM engineering, currently based in New York and holding dual graduate degrees with 4.0 GPAs from the University of Wisconsin–Madison. Their research and industry work centers on human-AI alignment, LLM-agent networks, and scalable post-training methods (SFT, DPO, RL-based fine-tuning) that have demonstrably improved multi-step reasoning and recommendation performance in production-scale systems. At PayPal they led distributed GPU training pipelines and applied RAG and multi-agent LLM architectures to boost recommendation precision and reduce inference cost, and previously developed graph-based fraud and viral marketing models on graphs with tens of millions of nodes. Academically they publish at ACL, EMNLP findings and NeurIPS workshops on evolving domain adaptation and simulating opinion dynamics with LLM agents, reflecting a rare combination of cognitive-science rigor and engineering impact. Beyond models, they’ve automated large neuroimaging pipelines and built robotics social-cognition modules, showing an interdisciplinary knack for turning behavioral insights into scalable ML systems. Their work is notable for treating LLMs as socio-cognitive agents—using networks of models to study and improve collective human-AI behavior rather than optimizing single-model metrics alone.
9 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy - PhD Psychology (Area: Cognitive Science), Doctor of Philosophy - PhD Psychology (Area: Cognitive Science) at University of Wisconsin-Madison
Bachelor of Science Psychology (Cognitive Science and Neuroscience Program), Bachelor of Science Psychology (Cognitive Science and Neuroscience Program) at National Taiwan University
Contributions:3 releases, 129 commits, 13 PRs in 3 years 4 months
brainneuroimagingaalmricoordinate
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.