Raafat Feki is an HPC Performance Engineer with 11 years of experience blending systems research and production software, currently optimizing high-performance storage and I/O at Cornelis Networks. He holds a PhD in Computer Science from the University of Houston and has a strong track record in parallel I/O, collective file operations, and GPU storage integration—contributing backend performance fixes to the widely used Open MPI project. His background spans academic teaching, GPU-tuned linear algebra libraries, and industrial R&D, giving him both theoretical rigor and hands-on optimization skills. Notably, he redesigned I/O stacks to support NVIDIA GPUDirect Storage and refactored Open MPI collective I/O to reduce allocations and enable larger offsets, showing a knack for low-level performance wins that scale.
11 years of coding experience
7 years of employment as a software developer
Associate's Degree, Mathematics and Physics., Associate's Degree, Mathematics and Physics. at Preparatory Institute for Engineering Studies - Sfax (IPEIS)
Baccalaureate degree, Mathematics, Highest honors, Baccalaureate degree, Mathematics, Highest honors at Pioneer School of Sfax
Doctor of Philosophy - PhD, Computer Science, GPA: 3.945, Doctor of Philosophy - PhD, Computer Science, GPA: 3.945 at University of Houston
Engineer's degree, Computer Science, GPA: 3.8, Engineer's degree, Computer Science, GPA: 3.8 at National school of computer science (ENSI)
5, Music, 5, Music at Conservatory of Music (Conservatoire Régionale De Musique De Sfax)
Contributions:16 reviews, 21 commits, 13 PRs in 2 years 10 months
Contributions summary:Raafat primarily contributes to the Open MPI project by modifying the file collective I/O operation code, specifically within the `fcoll/dynamic_gen2` and `fcoll/vulcan` directories. They are focused on optimizing the performance and efficiency of the collective file write operations. Their work includes refactoring the code to use derived datatypes for sending data, fixing warnings identified by Coverity, and minimizing memory allocations. They also made changes to improve code formatting and support larger offsets.
Contributions:27 pushes, 16 branches in 6 years 2 months
javampiswig
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.
Request Free Trial
Raafat Feki - HPC Performance Engineer at Cornelis Networks