Madhushree Kumari

Title of the Talk: Autonomous Database Sharding: Using Machine Learning for Dynamic Rebalancing

Abstract of the Talk: Large databases often split data into smaller pieces called “shards” across different servers. Normally, this is done using fixed rules, which can cause problems when some servers get overloaded with too much traffic while others stay mostly empty. Fixing this usually requires slow, manual work by database administrators.

In this talk, we will explore how machine learning can make databases smarter and self-managing. We will look at how AI can track database activity in real time, predict traffic spikes before they happen, and automatically move data to keep servers balanced and fast. We will also discuss the main technical challenges and what the future holds for smart, self-healing database systems.

Bio:
Madhushree Kumari is a Staff Software Engineer at Visa with over 13 years of experience architecting scalable distributed systems. She currently leads a global team within Visa’s FX trading platform, specializing in resilient microservices, cloud-native solutions and AI. Her career spans impactful engineering roles at SAP, McAfee, IBM, and FIS.

An alumna of Georgia Tech and currently pursuing AI studies at Stanford, Madhushree is a recognized leader in the tech community. She serves as Chair for IEEE Young Professionals Atlanta, Secretary for the IEEE Atlanta Section, and is a Founding Member of WomenTech Network. Her contributions earned her the 2025 WomenTech Global Silver Award for Software Engineering Leader of the Year. Beyond engineering, she is a frequent keynote speaker, peer reviewer, and mentor for organizations including AnitaB.org and Georgia Tech.