I am a Senior Research Staff Member at the IBM T. J. Watson Research Center (NY, USA) involved in R+D work in the areas of heterogeneous systems, embedded designs, and edge AI computing. I made contributions to the field with top-tier publications (e.g. in MICRO and PACT) and have been appointed as General Chair of the 2024 International Symposium on Computer Architecture (ISCA), the flagship conference in computer architecture. I have more than 20 issued US patents and others pending. I also have several peer-reviewed publications and a co-guest edited book on rugged embedded systems. As a Research Staff Member at IBM, I have helped spawn off the new area of cloud-backed, swarm-AI edge cognition systems with application to autonomous/connected vehicles, drones, wearable devices, mobile health monitoring and cyber-physical systems, among others. Today, I am the CTO of the DARPA-funded EPOCHS project (currently ongoing), and I previously contributed to the DARPA PERFECT program. I have also served on multiple conference technical program committees.
Grade: Summa Cum Laude (“with highest honor")
Conduct highly qualified research activities within the Efficient and Resilient Systems group (led by Dr. Pradip Bose). Involved in R+D work in the areas of
heterogeneous systems, embedded designs, and edge AI computing. As a Research Staff Member at IBM, I have helped spawn off the new area of cloud-backed, swarm-AI
edge cognition systems with application to autonomous/connected vehicles, drones, wearable devices, mobile health monitoring and cyber-physical systems, among
others. I am the CTO of the DARPA-funded EPOCHS project (currently ongoing), and I previously contributed to the DARPA PERFECT program.
In the past, I also conducted R+D work in support of IBM System p and Data Centric Systems, with primary focus on power-aware computer architectures and associated
system solutions. Issued several invention patents and published several papers in top-tier conferences and journals.
Developed techniques to reduce chip power consumption in multicore/manycore chips for multi-threaded applications, exploiting core folding, frequency/voltage scaling and low-power (“sleep”) modes. Developed an infrastructure to evaluate these power-reduction techniques in real POWER processors. Worked on applications for Unmanned Aerial Vehicles (UAVs) to adapt them to operate under harsh conditions such as high performance requirements and tight power budgets. Worked and developed the processor-in-regfile (PIR) technique which consists of computation logic embedded in a register file to accelerate parallel programs.
Devised a last-level cache (LLC) organization to improve bandwidth in multicore/manycore chips. In contrast to state-of-the-art LLC designs, this organization avoids data replication and, hence, does not require keeping data coherent. This results in better hit rates and higher bandwidth compared to a conventional (data coherent) LLC. Developed TaskSim, a modular CMP-architecture performance modeling tool.
Devised a first-level cache (L1) optimized for SMT-enabled chip multiprocessors. Developed a performance model to study first-level caches in SMT-enabled chip multiprocessors.
Introductory and advanced undergraduate courses on computer architectures and programming.