Summary
Sergio Perez is an AI Solution Architect with 10 years of experience building math-driven, GPU-accelerated AI systems across academia and industry, currently translating NVIDIA’s hardware advances into production-ready solutions for researchers and enterprises. He combines a PhD in data-driven fluid dynamics from Imperial College London with hands-on ML engineering at Graphcore and Amazon, optimizing LLMs, vision models and GNNs for novel accelerators and multi-device clusters. Sergio specializes in model compression and precision strategies (FP8, NVFP4), throughput optimization, and end-to-end proof-of-concept delivery that bridges research performance gains to real-world deployments. His background in high-performance scientific computing and teaching (including Coursera courses with 15K+ students) gives him a rare blend of deep numerical methods expertise and practical developer-facing communication. Colleagues rely on him not only for technical strategy but also for making complex HPC and ML tooling accessible to cross-disciplinary teams.
10 years of coding experience