Mohammad Ewais
Ml Hpc Architect @AMD
Signup · Get unlimited contacts
WORK HISTORY
Ml Hpc Architect @AMD
Markham, ON, CA
Successfully created/generated workloads for new target applications and optimizing these workloads to fit our line of GPUs.• Analyzed to find performance bottlenecks, fix performance bugs, or suggest new architecture features.• Conducted theoretical analysis and return on investment calculations for suggested features, ranging from spreadsheet‑based calculations to sophisticated python models, and taking into consideration real bandwidth and power numbers of previous generations.• Modeled target features using our in‑house simulator to get working proof of concepts and confirm return on investment calculations.• Collaborated with others to deliver clear and comprehensive feature specifications, ensuring seamless handoff to design and verification teams.
EDUCATION
University of Toronto
Doctor of Philosophy - PhD, Computer Engineering
Alexandria University
Bachelor's degree, Electronics and Communications Engineering
The University of British Columbia
Master of Applied Science, Computer Engineering
SKILLS
ABOUT MOHAMMAD EWAIS
I\'m currently doing my PhD in Computer Engineering at the University of Toronto, I also work as a ML/HPC Architect at AMD. In my PhD, I am currently focused on exploring disaggregated memory architecture for datacenter. At AMD, I work on designing/architecting a line of datacenter GPUs to better support machine learning applicationsI have a strong passion for computer architecture, digital design, and CAD tools, and a comprehensive background spanning hardware and software. I have accumulated over 4 years of experience in ASIC/FPGA design. I have also conducted research in computer architecture and memory hierarchy optimizations for 5+ years, gaining experience with multiple computer architecture simulators, as well as building my own from scratch. Additionally, I have contributed to LLVM-based compiler development for over 1 year.
This profile is compiled from publicly available professional sources. Unifers is not affiliated with or endorsed by LinkedIn. Request removal of this profile.