AI-Powered Image Generation
Pioneered transformer architecture adapting GPT-2 for pixel sequence prediction, achieving 99% accuracy on CIFAR-10.
Impact: Outperformed CNNs by 12% margin while requiring 60% less labeled data.
Software Engineer
Kamal Saad is a Software Engineer specializing in developing advanced systems, 3D applications, and AI-powered solutions. With expertise in C++, C#, and Python, he excels in distributed systems, game development, and machine learning, driving innovation and efficiency.
Expertise in designing and implementing high-performance 3D applications and game systems, including real-time networking, spatial partitioning, and VR visualization tools.
Specialization in developing AI-powered solutions, including generative models for image and narrative creation, and fine-tuning language models like GPT-2.
Proficient in building scalable backend services and distributed systems, focusing on robust architecture, REST API design, and database management.
Pioneered transformer architecture adapting GPT-2 for pixel sequence prediction, achieving 99% accuracy on CIFAR-10.
Impact: Outperformed CNNs by 12% margin while requiring 60% less labeled data.
Designed and implemented a 3D storytelling platform using Unreal Engine 4 and GPT-2.
Impact: Improved narrative coherence by 40% through fine-tuned language models.
Developed a C++ plugin for automated animation conversion.
Impact: Processed 30,000+ BVH files and saved 2,000+ manual hours.
Spearheaded development of a low-latency networking layer.
Impact: Reduced multiplayer lag by 22% through prediction algorithms.
Joined Microsoft (Bing Team) as a Software Engineer.
Completed Bachelor of Computer and Information Sciences from Ain Shams University.
Co-founded Odelay Studio, serving as Lead Software Engineer & CTO.
Achieved Top 15% in ACM-ICPC Egypt Regional.
Awarded 1st Place in FCIS Mini Game Jam.
Led OSC Game Committee, organizing events for 500+ participants and conducting workshops.