Evangelos Georganas
Principal AI Engineer at Intel Labs
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolEvangelos Georganas is a Principal AI Engineer based in Berkeley with 13 years of experience bridging academic research and production-grade AI systems. He holds a PhD in Computer Science from UC Berkeley and spent a decade at Intel Labs advancing performance-focused ML tooling before stepping into a principal engineering role. His work blends low-level performance engineering—evidenced by contributions to the libxsmm project optimizing dense and sparse matrix kernels—with research expertise from postdoc and graduate research positions. Evangelos specializes in squeezing hardware efficiency out of deep learning primitives and creating robust binary descriptors and norm-checking logic that reveal a deep understanding of hardware-software co-design. Colleagues rely on him for translating cutting-edge research into highly optimized, deployable components for large-scale AI workloads.
13 years of coding experience
14 years of employment as a software developer
PhD, Computer Science, PhD, Computer Science at UC BerkeleyElectrical and Computer Engineering, Electrical and Computer Engineering at National Technical University of Athens
Greek, English, German
Github Skills (7)
c1710
mat10
matrix10
c1110
performance-optimization10
arm9
deep-learning8
Programming languages (4)
Github contributions (5)
libxsmm/libxsmm
Mar 2017 - Jan 2023
Library for specialized dense and sparse matrix operations, and deep learning primitives.
Role in this project:
Back-end Developer & Performance Engineer
Back-end Developer & Performance EngineerContributions:168 reviews, 1600 commits, 57 PRs in 5 years 11 months
Contributions summary:Evangelos primarily focused on low-level optimization and performance improvements within the context of specialized matrix operations and deep learning primitives. Their contributions include optimizing code within the generator files for specialized dense and sparse matrix operations. They implemented better norm checking and created binary descriptors, showing a strong understanding of the underlying hardware.
bfloat16avxsimdlapackavx512
libxsmm/parlooper
Jun 2022 - Jan 2023
PArallelLOOPgEneratoR: Threaded Loops Code Generation Infrastructure targeting Tensor Contraction Applications such as GEMMs, Convolutions and Fused Deep Learning Primitives
Contributions:22 reviews, 101 commits, 45 PRs in 7 months
code-generationloopstensor-contractionconvolutionsdeep-learning
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.