Geoffrey Wenger is a Principal Software Engineer based in San Diego with a decade of hands-on experience building high-performance deep learning systems from embedded real-time inference engines to custom training accelerators. He has driven GPU and accelerator software across industry leaders—Qualcomm, Intel, Cerebras, and d-Matrix—often delivering runtime optimizations, FP16/math advances, and compiler-level techniques like precompilation, tensor lifetime analysis, and heterogeneous op segmentation. Geoffrey pairs low-level kernel tuning and threading expertise with higher-level compiler/runtime design, improving both performance and numerical verification for AI workloads. Comfortable across mobile GPUs, datacenter accelerators, and edge devices, he repeatedly translates research ideas into production features used in shipping SDKs and chips. His background in mathematics and computer science underpins a pragmatic approach to precision, pipelining, and memory efficiency that surfaces in unexpected places, such as doubling down on verification workflows to catch subtle numerical drift.
10 years of coding experience
23 years of employment as a software developer
BA Mathematics, BA Mathematics at Franklin & Marshall College
MS Computer Science, MS Computer Science at Rensselaer Polytechnic Institute
nGraph - open source C++ library, compiler and runtime for Deep Learning
Contributions:29 PRs, 63 pushes, 33 branches in 1 year 5 months
inference-enginecppc-librarydeep-learningtvm
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.