모바일 메뉴 닫기
 

커뮤니티

제목
[세미나] ENCO: Deploying Production-Scale Engineering Copilots / Yiwen Zhu (Principal Scientist at Microsoft’s GSL)
작성자
첨단컴퓨팅학부
작성일
2025.10.17
최종수정일
2025.10.17
분류
세미나
링크URL
게시글 내용



일시: 2025. 11. 14. (금요일), 오후 2시 ~ 오후 3시
장소: 제4공학관 D604호

Speaker: Yiwen Zhu, Ph.D. (Principal Scientist at Microsoft’s Gray Systems Lab (GSL))


Title: ENCO: Deploying Production-Scale Engineering Copilots


Abstract:

Software engineers often face significant hurdles when navigating disparate sources of documentation and telemetry data—ranging from troubleshooting guides (TSGs) and incident reports to code repositories and internally developed tools maintained by different stakeholders. On-call responsibilities further compound this challenge: resolving incidents under strict time constraints becomes particularly difficult when legacy sources are obscure or fragmented.


To address these inefficiencies, we introduced ENCO—a comprehensive framework for developing, deploying, and managing enterprise-grade copilots that enhance productivity in engineering workflows. ENCO integrates customized retrieval-augmented generation (RAG) algorithms that not only surface relevant information from diverse data sources but also identify and invoke the most appropriate skills to resolve user queries. Building on this foundation, we are extending the framework with SuperDRI, which aims for fully end-to-end automation of incident mitigation by leveraging learned troubleshooting graphs.


Since its launch in September 2023, ENCO has demonstrated strong adoption and impact, supporting tens of thousands of interactions and engaging more than one thousand monthly active users (MAU) across dozens of organizations within the company.


Bio:

Yiwen Zhu is a Principal Scientist at Microsoft’s Gray Systems Lab (GSL). Her research interests center on the vision of autonomous cloud systems, utilizing machine learning, statistical inference, and operation research techniques. Her work on data-driven optimization for cloud services has delivered tens of millions of dollars in annual cost savings and underpins efforts toward autonomous data services on Azure. Additionally, Yiwen leads the development of an internal large language model (LLM) application aimed at improving workflows for software engineers.