Muhammad Haseeb

HPC Infrastructure & Performance Postdoc
NERSC, Berkeley Lab
1 Cyclotron Road,  Bldg. 59,
Berkeley, CA 94720, USA

Email: mhsb123@gmail.com

Professional Summary

Experienced software engineer and researcher developing cutting-edge GPU-accelerated infrastructure for HPC, AI and scientific computing. Adept in modern C++, python, accelerator programming, distributed systems, programming models, data structures and algorithms, performance optimization, neural networks, and build systems.

Work Experience

2023 - Present | HPC Infrastructure & Performance Postdoc | NERSC, Berkeley Lab
Develop cutting-edge GPU accelerated scientific software using new technologies in Programming Models (MPI, CUDA, SYCL, Kokkos, OpenMP-offload, AMReX), and C++ (stdexec, parSTL). Model GPU-GPU communications in HPC applications over Perlmutter supercomputer interconnects.

2021 (Summer) | Application Performance Intern | NERSC, Berkeley Lab
Developed and evaluated DPC++ (SYCL) and Python-based CPU-GPU software for ADEPT sequence alignment kernels.

2020 (Summer) | Application Performance Intern | NERSC, Berkeley Lab
Contributed to the development of dynamic, MPI, Python instrumentation, data analytics, and continuous integration (CI) features for Timemory

2018 - 2023 | Graduate Research Assistant | Florida International University
Developed parallel algorithms, data structures, and GPU kernels to scalably accelerate computational proteomics algorithms by > 40x on modern supercomputers.

2015 - 2017 | Senior Software Engineer | Mentor Graphics Corp (now Siemens EDA)
Contributed to the development of system partitioning, low-level drivers, remote life-cycle management, inter-processor communication, and tracing features for the Mentor Embedded Nucleus RTOS and the Mentor Embedded Multicore Framework (MEMF).

Education

2018 - 2023 | Ph.D. Computer Science | Florida International University

2011 - 2015 | B.Sc. Electrical Engineering | University of Engineering and Technology, Lahore

Skills & Interests

C++ Evolution Programming Models • Accelerator Programming • Experimental TechnologiesAI Software Infrastructure Performance Engineering Algorithms & Data Structures • Software Integration & Build Systems

Software Development

C++26 • Python • CUDA • MPI • DPC++/SYCL • Kokkos • OpenMP • CMake • Git • Thrust • AMReX • PyBind11 • Nsight tools • Intel tools • NCCL • NVSHMEM • Embedded C • TorchTensorFlow • Bash • C++ Evolution.

Select Recent Publications (Google Scholar)

2023 | Muhammad Haseeb, and Fahad Saeed. "GPU-Acceleration of the Distributed-Memory Database Peptide Search of Mass Spectrometry Data.", Nature Scientific Reports, 13, 18713 (2023). https://doi.org/10.1038/s41598-023-43033-w

2023 | Nan Ding, Muhammad Haseeb, Taylor Groves, and Samuel Williams. "Evaluating Performance of One-sided MPI communication on GPUs and GPUs.", 2023 International Workshop on Performance, Portability and Productivity in HPC (P3HPC - SC '23), IEEE, 2023.

2023 | LeAnn Lindsey, Muhammad Haseeb, Hari Sunday, and Muaaz Awan. "TANGO: A GPU-Optimized Traceback Approach for Sequence Alignment Algorithms.", 13th International Workshop on Irregular Applications: Architectures and Algorithms (IA^3 - SC '23), IEEE, 2023.

2021 | Muhammad Haseeb, and Fahad Saeed. "High performance computing framework for tera-scale database search of mass spectrometry data." Nature Computational Science, Volume no. 1, Issue no. 8 (2021): pp. 550-561.

Patents

2022 | Muhammad Haseeb, and Fahad Saeed. "Systems and Methods for Peptide Identification." U.S. Patent 11,309,061 issued April 19, 2022.

2020 | Fahad Saeed, and Muhammad Haseeb. "Methods and systems for compressing data." U.S. Patent 10,810,180, issued October 20, 2020.

Awards

2018 - 2023 | Tuition Waiver for PhD studies, School of Computing and Information Sciences, FIU

2022 & 2021 | Patent Award, Office of Research and Economic Development, FIU

2021| Best Graduate Student Research Award, School of Computing and Information Sciences, FIU

2016 - 2017 | Role Model Software Engineer, Mentor Graphics Corporation (now Siemens EDA)