I am a researcher at VMware Research Group, where the principle focus of my work is the intersection of external memory theory and storage systems. I am especially interested in key-value stores, file systems, filters and other systems built on a strong theoretical foundation.
My recent publications include work on key-value stores for fast storage, file system aging, external memory hash tables and the analysis of storage optimization heuristics.
I am the research lead for the SplinterDB project. SplinterDB is a general purpose key-value store built using a new data structure, the size-tiered $B^\varepsilon$-tree. SplinterDB is designed for outstanding performance on NVMe and other fast storage hardware, under tough conditions, such as strict memory limits, small key-value pairs and limits on CPU utilization.
I am a member of the BetrFS team, which built and maintains the BetrFS prototype file system. BetrFS is designed around the principles of write-optimization, and is built using variants of $B^\varepsilon$-trees.