I lead CrocodileDB, a next-generation database that helps users ingest and process the fast-changing data. CrocodileDB avoids the trap of unnecessary resource consumption and expedites result delivery. It integrates workload prioritization, execution policies, query scheduling, and data loading.
I extended transaction processing theory to study how large-scale optimization behaves in the asynchronous environment, especially when/how asynchronization endangers convergence.
SIGOPS Operating Systems Review
My dissertation focuses on efficient and scalable graph computing systems that preserve the correctness when necessary, and strike at the balance between performance and accuracy when approximations are preferred. I also designed graph analytical algorithms.
Frontiers of Computer Science