Summary
Dezhi Hong is an applied scientist with 12 years of experience building intelligent systems that integrate large foundation models, machine learning, and control theory to reason over multimodal data such as text, time series, and speech. He has transitioned research into product impact across academia and industry—from postdoctoral research and research faculty at UC San Diego to applied science roles at Amazon (leading LLM post-training for tool use and sustainability initiatives) and now driving agentic workflows for Copilot at Microsoft. His work enables adaptive home automation, context-aware voice assistants, and intelligent wearables focused on well-being, productivity, and sustainability, blending rigorous theory with practical deployment. Dezhi’s profile reflects deep expertise in multimodal representation and control-aware ML, with a publication record available on Google Scholar that signals continued academic engagement alongside product-focused innovation. Based in the Greater Seattle Area, he brings a rare combination of control-theoretic insight and large-model engineering to scale contextual, proactive interfaces.
12 years of coding experience
10 years of employment as a software developer
Bachelor of Science (B.S.), Electrical, Electronics and Communications Engineering, Bachelor of Science (B.S.), Electrical, Electronics and Communications Engineering at Beijing University of Posts and Telecommunications
Doctor of Philosophy (PhD), Computer Science, Doctor of Philosophy (PhD), Computer Science at University of Virginia
Chinese, English