Mohamed Deraz Nasr

ML researcher + engineer interested in embodied/physical AI, robotics, neural graphics, inference & distributed systems.

Education

Georgia Institute of Technology

M.Sc. in Computer Science· GPA 3.9 · Jan 2026 – Dec 2027

University of Ottawa

B.Sc. in Software Engineering · GPA 3.7 · Sep 2020 – May 2025

Affinity Map: Few-Shot Protein Family Classification via Prototypical Networks

Zenodo · 2026

Georgia Institute of Technology · Atlanta, Georgia · Jan 2026 – Present

Machine Learning Researcher

University of Ottawa · Ottawa, Canada · Sep 2024 – Dec 2024

Research Engineer Intern

Shopify · Ottawa, Canada · Jan 2025 – Apr 2025

Machine Learning Engineer Intern

March Networks · Ottawa, Canada · Jan 2023 – Apr 2023

Systems Engineer Intern

· An SE(3)-equivariant 3D Gaussian world model for scene understanding, dynamics, and robot manipulation.

· C++/CUDA inference engine for diffusion models targetting 10x speedup over PyTorch via INT8 quantization + persistent kernels.

· A 3D protein backbone diffusion system using equivariant modeling and AlphaFold-based plausibility checks.

· A F1 strategy simulator trained with PPO & behavioral cloning that exceeds baseline performance.

UOBionics

Allonstride Exoskeleton · Embedded Developer · Jan 2024 – Dec 2024