Summary
Seunghak Yu is a Principal Scientist with nine years of experience building production-grade conversational AI and context-aware advertising systems, currently leading NAVER’s Agentic Ads initiative from Bellevue, WA. He has driven end-to-end launches of generative AI products (including Cue:) and designed LLM-based answer generation and verification platforms that balance monetization, user experience, and answer quality. His background spans industrial deployments at Amazon Alexa and Samsung—where his neural language model shipped on over 100 million devices—and academic research at MIT CSAIL on conversational access and misinformation detection. Seunghak combines rigorous ML research (PhD in IT Convergence) with hands-on system architecture, orchestration, and evaluation pipelines for online/offline ad decisioning. He’s notable for integrating doubly robust estimators and OPE methods into real-world routing and ad systems to reduce bias and accelerate iteration. Fluent at the intersection of research and product, he excels at turning advanced NLP and causal-evaluation techniques into scalable commercial features.
9 years of coding experience
12 years of employment as a software developer
Doctor of Philosophy (PhD), IT Convergence (Machine Learning), Doctor of Philosophy (PhD), IT Convergence (Machine Learning) at Korea University
English, Japanese, Korean