Irina Yu

About Irina Yu

Focus on developing cutting-edge LLM-driven AI systems and machine learning solutions that enhance customer experiences. Since 2021, I\'ve spearheaded innovations in defect reduction and personalization systems, impacting tens of millions of customers globally.Currently leading the technical development of Amazon Customer Service\'s foundational LLM-driven agentic system, transforming theoretical concepts into scalable operational frameworks. This groundbreaking work sets the foundation for future AI innovations at Amazon.Beyond my core responsibilities, I\'m passionate about mentoring the next generation of AI practitioners, conducting research, and sharing insights through publications and workshops. I thrive on designing and implementing innovative ML systems that solve complex real-world challenges.Areas of expertise: Large Language Models, Agentic AI, Personalized AI Systems, AI Agent, Machine Learning, Technical Leadership

Experience

1. Applied Ml Scientist & Tech Lead

   Amazon Web Services (AWS)

   Oct 2021 — Present · Seattle, WA, US

   Leading science design, prototyping and development of the foundational LLM-driven agentic system for Amazon\'s AI chatbot. This work transformed an initially ambiguous concept into a fully operational framework, which now serves as a foundation for future innovations in Amazon\'s customer feedback analysis and insight generation systems.Developing AI-driven solutions improving the experience for tens of millions of Amazon customers, including automated customer feedback analysis, intelligent insight generation, contextual theme extraction, explainable summarization, and LLM-based natural language interaction for customer feedback understanding.

Education

• Shanghai University of Finance and Economics

  Bachelor's degree, Information and Computational Science

• Temple University

  Doctor of Philosophy - PhD, Computational Science and Applied Mathematics

  2014 — 2020