Mohammad Nazmus Sakib is a Research Computer Scientist at USC Information Sciences Institute where he advances next-generation computing architectures. His expertise spans processing-in-memory (PiM) design, FPGA-based system emulation, and hardware acceleration for machine learning applications.

Sakib's research journey bridges the gap between computational logic and memory architectures through innovative design methodologies. He has made significant contributions to PiMulator, an FPGA emulation framework that substantially accelerates the evaluation of PiM architectures compared to traditional simulation methods. His work on MEMulator further extends this approach by enabling detailed design space exploration across various memory technologies including DDRx, HBM, HMC, and STT-MRAM.

In the realm of temporal computing, Sakib collaborated with NIST and the University of Maryland to develop non-volatile magnetic racetrack memory for race logic applications. This research led to the development of temporal PiM architectures that significantly outperform conventional classifiers through specialized hardware-algorithm co-design.

During his time at Qualcomm, Sakib pioneered machine learning-based CPU power modeling techniques for Snapdragon chipsets. By creating innovative correlations between performance monitoring data and power consumption, he developed predictive models that enhance accuracy over traditional approaches.

Through his ongoing research, Sakib continues to develop innovative solutions at the intersection of memory systems, hardware acceleration, and efficient computing architectures, pushing the boundaries of what's possible in modern computing systems.

UVA Engineering Endowed Fellowship
Co-authored a Best Paper Award winner at ISQED 2024