Alberto Magni - Compiler Engineer at DeepMind

Alberto Magni

Compiler Engineer at DeepMind

United Kingdom

Join Prog.AI to see contacts

Summary

🤩

Rockstar

🎓

Top School

Alberto Magni is a compiler engineer with eight years of professional experience building high-performance tooling for ML and graphics at DeepMind, NVIDIA and Microsoft. He combines a strong research pedigree (PhD in Computer Science from the University of Edinburgh) with hands-on implementation experience spanning compiler backends, optimizations and runtime model handling. At Microsoft he contributed to ONNX Runtime—adding shape inference for operators and workarounds to bypass protobuf limits—demonstrating attention to both correctness and practical production constraints. His background includes multiple roles at NVIDIA (including OptiX compiler work) and a track record of improving C++ code quality and core functionality in cross-platform projects. Based in the UK, he brings a pragmatic blend of research-driven problem solving and production-grade engineering focused on performant ML inference and compilation pipelines.

7 years of coding experience

9 years of employment as a software developer

Doctor of Philosophy (Ph.D.), Computer Science, Doctor of Philosophy (Ph.D.), Computer Science at The University of Edinburgh

Master Degree, Computer Science, Master Degree, Computer Science at University of Illinois at Chicago

Master Degree, Computer Engineering, Master Degree, Computer Engineering at Politecnico di Milano

Italian, English

Github Skills (9)

machine-learning10

c-language10

onnx10

cprogramming-language10

deep-learning9

hardware-acceleration5

pytorch4

python4

tensorflow4

Programming languages (1)

C++

Github contributions (5)

microsoft/onnxruntime

Jul 2020 - Mar 2021

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator

Role in this project:

Backend Developer

Contributions:14 reviews, 6 commits, 9 PRs in 7 months

Contributions summary:Alberto contributed to the core functionality of the ONNX Runtime by addressing code quality issues, such as fixing indentation and adding braces for better readability in the C++ code. They also implemented and added shape inference support for new operators like SoftmaxCrossEntropy and LayerNormalizationGrad. Furthermore, the user made changes related to saving and loading models, including enabling the use of external files for initializers to bypass the 2GB protobuf limit.

runtimetrainingtensorflowai-frameworkaccelerator

alberto-magni/onnxruntime

Oct 2020 - Feb 2021

ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator

Contributions:11 pushes, 7 branches in 4 months

pytorchdeep-learningruntimemachine-learningonnx

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.

Request Free Trial