Mahmoud Ellithy

Technical leader with 15 years in production VFX, including a decade building tools, workflows, and pipeline infrastructure, and the last 5 years focused on ML/AI integration at production scale. I work at the intersection of software engineering, computer vision, and creative pipelines—deploying AI systems that operate reliably under real production constraints. I specialize in identifying bottlenecks and architecting solutions that become production-critical infrastructure. My work spans end-to-end system design, including large-scale GPU orchestration platforms, AI generation and inference services, USD-centric asset management systems, and review pipelines integrating real-time tooling with studio databases. These systems routinely support hundreds of concurrent operations, multi-site teams, and tight delivery schedules. My technical background covers deep learning (PyTorch, TensorFlow), computer vision, and applied ML, alongside strong foundations in system architecture, GPU computing, and API design. I have hands-on experience deploying models into production environments, designing scalable data flows, and building robust async and distributed systems in Python and C++, with a focus on maintainability and performance. Across my career, I’ve delivered ML inference systems adopted at studio scale, built pipeline stacks enabling delivery of hundreds of shots, and developed artist-facing tools that significantly improved iteration speed and final quality. I’m comfortable leading cross-functional initiatives, mentoring engineers and artists, and translating complex technical ideas into practical, usable workflows. My core strength is bridging deep technical expertise with production reality—building systems that solve hard problems, scale under pressure, and are genuinely embraced by the people using them.

Role with the VES

Member