Tony’s work on a dynamically tunable key-value storage engine (TurtleKV) was accepted to VLDB 2026! This paper explores how to utilize the write-memory dimensions of the RUM space to dynamically improve read or write performance. You can read the ArXiV version here.