A compact map of the systems themes I care about: CPU/GPU platforms, CXL memory systems, AI workloads, and energy-harvesting devices.
Studying where computation, memory, and data movement interact across server-class systems.
Persistence, resilience, and correctness as first-class design goals for emerging memory and compute systems.
Connecting system-level design with AI workloads, accelerators, memory behavior, and efficient execution.
Exploring computation that adapts to intermittent power, constrained energy budgets, and intelligent sensing.
The workflow: measure real behavior, identify bottlenecks, co-design software/hardware mechanisms, and validate the effect across workloads.