Skip to content
AI LearningAtlas

Yuanquan Hu — Author

3 min read

Properties2
descriptionMeet Yuanquan Hu, author of AI Learning Atlas and a senior AI product manager building trustworthy agent, search, grounding, and evaluation systems.
tagsauthor, ai-product-management, agents, search, grounding, evaluation

Author · AI Learning Atlas

Yuanquan Hu

Senior AI Product Manager focused on AI agents, search and grounding, context intelligence, evaluation, reliability, and trustworthy AI product systems.

I build AI products that help people move from open-ended questions to useful, verifiable outcomes. Since joining Microsoft in 2019, my work has spanned Excel Agent, Excel Copilot, Microsoft Forms, and Cortana, with a consistent focus on making capable systems understandable, measurable, and ready for responsible use.

Focus

What I work on

AI search and knowledge systems Agents and context engineering RAG, grounding, and provenance Evaluation and release gates Model routing and cost-performance decisions Enterprise AI product strategy Hardware and software product delivery Cross-functional product leadership

Experience

Microsoft journey

  1. 2025–present

    Excel Agent

    Product leadership for search, grounding, and context intelligence, including tiered retrieval, sub-agent, and multi-model strategies. I focus on release readiness, reliability, provenance, quality, and cost-aware system decisions.

  2. 2024–2025

    Excel Copilot

    Launched AI-assisted content-creation experiences, multi-turn LLM workflows, grounding and RAG experiments, and evaluation systems while coordinating engineering, science, design, annotation, and partner teams.

  3. 2020–2024

    Microsoft Forms

    Launched Copilot and AI-assisted survey and learning workflows across commercial, consumer, and education contexts, established quality feedback loops, and led broader product growth work.

  4. 2019–2020

    Cortana and NLP API adoption

    Localized multilingual conversational capabilities and defined quality and latency expectations for partner integrations.

Principles

How I build AI products

01

Ground answers in sources

Useful AI should show where its claims come from and make evidence easy to inspect.

02

Measure quality

Evaluation, feedback loops, and release gates turn product intuition into dependable decisions.

03

Keep provenance transparent

People should understand which context, tools, and models shaped an answer.

04

Preserve user control

Responsible rollout pairs system capability with clear choices, predictable behavior, and a path to correction.

中文简介

关于作者

Yuanquan Hu 是一名专注于 AI 智能体、AI 搜索、知识增强、上下文智能与评估体系的资深产品经理。他在微软负责过 Excel Agent、Excel Copilot、Microsoft Forms 和 Cortana 相关产品工作。他重视有来源依据的回答、可衡量的质量、透明的溯源、用户控制和负责任的产品发布。

Connect

Keep in touch

Follow my public work and learning notes through these profiles.

LinkedIn
Xiaohongshu ID 4748526758
Explore connectionsGraph and backlinks

Graph View