Koray Kavukcuoglu
Chief AI Architect at Google
London, England, United States
Summary
🤩
Rockstar🎓
Top SchoolKoray Kavukcuoglu is a seasoned AI leader with 15+ years of experience bridging cutting-edge research and large-scale productization, currently serving as Chief AI Architect at Google and CTO of DeepMind. He progressed through DeepMind’s research leadership—from research scientist to VP of Research—driving influential advances in machine learning and production AI systems. With deep academic roots (PhD and MS in Computer Science from NYU) and an earlier background in aerospace engineering, he blends rigorous theory with systems-level engineering. An active contributor to foundational ML tooling, he has made substantive low-level contributions to Torch (torch7) and PyTorch’s nn modules, including numerical and performance-focused fixes that underpin many modern models. Colleagues know him for pairing strategic vision with hands-on changes to core libraries—an uncommon mix that accelerates both research and deployment. Based in London, he brings a rare combination of long-term research impact and practical implementation expertise across academia and industry.
14 years of coding experience
12 years of employment as a software developer
PhD, Computer Science, PhD, Computer Science at New York UniversityMS, Aerospace Engineering, MS, Aerospace Engineering at Orta Doğu Teknik Üniversitesi / Middle East Technical University
Github Skills (17)
pytorch10
convolutional-neural-networks10
operation10
tensorrt10
machine-learning10
c1110
c1710
lua10
deep-learning10
tensorflow10
neural-network10
tensor10
testing9
linear-algebra9
lapack9
Programming languages (6)
Github contributions (5)
torch/torch7
Feb 2012 - Dec 2015
http://torch.ch
Role in this project:
Back-end Developer
Back-end DeveloperContributions:74 commits, 17 PRs, 22 pushes in 3 years 10 months
Contributions summary:Koray primarily contributed to the `torch/torch7` repository by implementing and extending core tensor operations. They introduced new functionalities such as `maskedSelect` and `atan2` methods, and added histogram calculation. Furthermore, the user addressed issues related to SVD calculations and matrix inverse. Their work involved modifications in both C and Lua files, demonstrating their expertise in the underlying implementation of the Torch library.
libtorchtorchc-plus-plustensorautograd
torch/nn
Feb 2012 - Oct 2015
Role in this project:ML Engineer
ML EngineerContributions:32 commits, 3 PRs, 3 pushes in 3 years 8 months
Contributions summary:Koray contributed to the PyTorch `nn` module by addressing bugs and implementing new features. Their work included correcting dimension checks in normalization layers, fixing a bug in the backward pass of the square function, and adding batch mode functionality to max pooling layers. They also added new NN classes and performed testing across several modules.
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.