Database Seminar: Yiwen Zhu (Microsoft): “Autonomous Data Management Systems”

Speaker: Yiwen Zhu (Gray Systems Lab, Microsoft)

Location: Soda 380

Date: October 18, 2023

Time: 11 AM – 12 PM PST

Title: “Autonomous Data Management Systems”

Abstract: In this talk, we delve into the nuanced world of leveraging machine learning (ML) to optimize big data systems, drawing insights from real-world projects and experiences. We begin by exploring the efficacy of leveraging ML, optimization, and data mining in projects like KEA and the Intelligent Pooling at Microsoft Azure Data. Those projects highlight the power of simplicity, demonstrating the great potential of ML applications.

This discussion provides a holistic view of the complete life cycle of ML in system projects, bridging the gap between mathematic models with the
complex realities in production environments.

Bio: Yiwen is a principal applied data scientist at Microsoft. She joined the Gray System Lab (GSL) research team in 2018 and is working on
data-driven optimization for cloud computing systems (e.g., Azure SQL, Cosmos, and Azure Synapse Analytics), achieving tens of millions of dollars’ cost savings per year. Her current research focuses on leveraging data science tools to improve system efficiency, such as machine learning, optimization, and copilot applications using LLM.