Research

My research explores runtime detection of microarchitectural attacks that exploit hardware features such as speculative execution, cache hierarchies, and branch prediction. I focus on detection mechanisms that can operate on existing systems, particularly those that cannot easily adopt new hardware or invasive software defenses.

Research threads

• Runtime detection of transient execution attacks
  Leveraging hardware performance counters to identify malicious behavior that has bypassed traditional mitigations.
• Performance counter semantics across architectures
  Understanding how counter behavior differs across Intel and ARM platforms and how these differences affect detection reliability.
• Data-driven detection pipelines
  Designing feature representations and machine learning models that operate on noisy, low-level hardware measurements.

These research threads are supported by the publications and projects described in the other sections of this Work area.