current

As of 2026-10-31

I’m finishing my master’s thesis on memory tiering for cloud workloads. CXL lets compute providers provision memory in a more efficient manner, but it only pays off if each workload lands on the right tier.

My main project is a VM placement harness that predicts a workload’s bandwidth and latency needs and uses that to decide how to split it across CXL and DRAM.

I am also a collaborator alongside a few other systems researchers working with a programming languages group on language-runtime support and performance analysis for memory-tier-aware programming.