Seo Taek Kong

I am a Ph.D. Candidate in the ECE department at the University of Illinois Urbana-Champaign, where I am fortunate to be advised by Prof. R. Srikant. My research develops the non-asymptotic analysis of discretized diffusion processes and applies it across three domains: stochastic gradient descent, reinforcement learning, and generative AI. By comparing stochastic approximation algorithms with discretized diffusions, I derive finite-time guarantees that characterize their behavior, and the same framework yields order-optimal noise schedules for sampling in generative diffusion models.

Alongside my academic work, I have built search engines in industry, developing an agentic video retrieval system at NVIDIA and a listwise ranking LLM at Amazon. Before my Ph.D., I spent three years as an AI Researcher and research team lead at VUNO Inc., where I developed deep learning models for medical imaging.