Olumuyiwa Oluwasanmi

I've led capacity planning and data center modeling for Meta AI, building supply, demand, and planning models that accelerated provisioning for machine learning workloads. My algorithms using data pipelines, operations research, and constraint programming reduced data center design time by 30%.

At Intel, I led performance modeling and simulation across the client product portfolio, developing software infrastructure and simulation frameworks for processor, graphics, and power optimization. I partnered with product and hardware architects, led technical teams, and mentored engineers across global teams.

Through Knobugsoft LLC, I deliver AI, performance optimization, capacity engineering, computational science, and mathematical modeling consulting. I also build high-performance C++23 and Python products using SIMD and GPU acceleration, and publish technical work on topics including DiffusionGemma, JEPA, Triton, and modern C++ error handling.

My background also includes doctoral research in Byzantine fault-tolerant distributed computing, enterprise financial systems development, and teaching computer architecture, algorithms, C++, Python, mathematics, and calculus.