Summary
Frank Mokadem is an ML compiler engineer with nine years of experience bridging research and production to make deep learning models leaner and faster for edge deployment. He has driven dramatic inference-cost reductions—up to 100x in product settings and 10x in academic projects—by combining neural architecture search, low-rank compression, and custom CUDA kernels for convolution factorization. Comfortable across C++, Python, CUDA and systems-level ML, he moves work from prototype to deployed tooling used by labeling and product teams. Based in Toronto with an MASc in AI-Systems Design, he excels at cross-team collaboration and formal, correctness-minded design, often surfacing practical evaluation metrics and automated configuration that aren’t obvious from model papers alone.
9 years of coding experience
5 years of employment as a software developer
Bachelor of Engineering - BE Computer Science, Bachelor of Engineering - BE Computer Science at ENSI - Ecole Nationale des Sciences de l'Informatique
Associate's degree Mathematics Physics and Chemistry, Associate's degree Mathematics Physics and Chemistry at Institue preparatoire aux etudes scientifiques et techniques
Master's of Applied Science AI - Systems Design Engineering, Master's of Applied Science AI - Systems Design Engineering at University of Waterloo
English, French, Arabic, German