Miltos Allamanis
Machine Learning Researcher
Greater Cambridge Area United Kingdom
Summary
🤩
Rockstar🎓
Top SchoolMiltos Allamanis is a machine learning researcher with 15 years of engineering and research experience, specializing in deep learning for structured data with a strong focus on real-world applications to source code. He holds a PhD from the University of Edinburgh and has held research roles at Microsoft (rising to Principal Researcher) and currently works at Google DeepMind, blending rigorous academic foundations with industry-scale impact. His contributions span both research and engineering: from implementing numerically robust softmax operations in the popular pytorch_scatter extension to pragmatic refactors across large open-source projects like libGDX and Thunderbird Android. Known for improving code quality and reproducibility as much as novel models, he bridges the gap between prototype research and production-ready tooling. Based in the Greater Cambridge area, he keeps an active public profile of publications and projects that reveal a continuing curiosity about practical program understanding and developer tools.
15 years of coding experience
7 years of employment as a software developer
Doctor of Philosophy (PhD), Doctor of Philosophy (PhD) at The University of EdinburghEngineering, Engineering at Aristoteleion Panepistimion ThessalonikisM.Phil., M.Phil. at University of Cambridge
Stackoverflow
89reputation
14kreached
1answer
4questions
Github Skills (28)
application-framework10
pytorch10
sendmail10
python10
send-email10
app-framework10
java10
e-mail10
android-developer10
javas10
android10
android-developers10
web-framework10
refactoring10
game-development10
Programming languages (7)
Github contributions (5)
rusty1s/pytorch_scatter
Nov 2019 - Oct 2021
PyTorch Extension Library of Optimized Scatter Operations
Role in this project:
ML Engineer
ML EngineerContributions:7 commits, 1 PR, 6 comments in 1 year 11 months
Contributions summary:Miltos implemented and refined numerical safe log-softmax and softmax operations within the `pytorch_scatter` library. They contributed to the `composite` module by adding these functions and addressed several bug fixes and formatting issues. The user's work involved modifying existing functions, testing, and improving the library's functionality for PyTorch, demonstrating a focus on extending and improving scatter operations.
pytorchoptimizedscattergathersegment
libgdx/libgdx
Feb 2014 - Feb 2014
Desktop/Android/HTML5/iOS Java game development framework
Role in this project:Backend Developer
Backend DeveloperContributions:5 commits in 1 day
Contributions summary:Miltos primarily contributed to code refactoring and maintenance within the libGDX game development framework. Their commits involve renaming variables and methods for improved code consistency and readability, as indicated by the "naturalize tool." These changes span multiple Java source files, touching upon core functionalities like vector math, XML parsing, and scene graph management. Their work appears to focus on improving the internal structure and maintainability of the existing codebase.
gamejava-game2dhtml5ios
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.