Yunfeng Hu

About Yunfeng Hu

As a Research Scientist at AWS, I have over six years of experience in applying mathematical and computational methods to solve real-world problems in cloud computing, machine learning, and artificial intelligence. I work with a diverse and collaborative team of researchers and engineers to develop innovative solutions that enhance the performance, scalability, and reliability of advanced models. Some of my achievements include designing, implementing, and evaluating cutting-edge algorithms and systems for optimizing resource allocation, reducing latency, and improving accuracy. I also leverage my teaching skills and experience to mentor and coach junior colleagues and interns, and to communicate complex concepts and results to various audiences. I have a PhD in Mathematics from Washington State University, where I conducted original research on measure theory, algebraic topology and optimization, and published several papers in reputable journals. I also have multiple certifications in R and Python, and I am fluent in English and Chinese. I am motivated by the mission of AWS to recommend fast, secure, and cost-effective cloud solutions to millions of customers around the world. I am always eager to learn new skills, explore new domains, and collaborate with different teams and partners. I believe that I can bring a unique and valuable perspective and experience to the AWS research community.

Experience

1. Applied Scientist Iii

   Amazon Web Services (AWS)

   Nov 2024 — Present · Seattle, WA, US

   I lead the design and deployment of large-scale GenAI and machine learning systems that turn complex behavioral and interaction data into actionable insights for product, UX, and leadership teams.My work includes building a GenAI-powered telemetry and adoption intelligence platform that serves as a centralized source of truth for understanding user behavior across AWS platforms. The system processes tens of millions of monthly interactions and uses ML and LLMs to classify user actions, segment journeys, and identify workflow bottlenecks. These insights directly inform product prioritization and UX improvements, reducing workflow complexity equivalent to hundreds of thousands of user hours per month.I also contribute to agentic AI platforms, where I own reasoning strategy, memory architecture, and evaluation design. This includes defining when to use tool-aware versus deep reasoning, designing multi-layer memory systems, and ensuring reliable citation and attribution across heterogeneous tools.In addition, I design GenAI-powered decision-support systems that combine deterministic logic with LLM-based synthesis to deliver accurate, explainable insights at scale. I partner closely with product, engineering, and business leaders to ensure these systems are trustworthy, actionable, and aligned with real user needs.

Education

• Central China Normal University

  Bachelor's degree, Mathematics

  2008 — 2012

• Washington State University

  Doctor of Philosophy, Mathematics

• Washington State University

  Doctor of Philosophy (PhD), Mathematics

  2013 — 2018

Skills

• Elasticsearch
• Higher Education
• Public Speaking
• Github
• Science
• Matlab
• Python
• Curriculum Development
• Statistics
• Lecturing
• University Teaching
• Microsoft Office
• Curriculum Design
• Research
• Mysql
• Teaching
• Geometric Measure Theory
• R
• Data Analysis
• Tutoring