Pedro Gonnet
Computer Scientist at google
Zurich, Zurich, Switzerland
Summary
🤩
RockstarPedro Gonnet is a computer scientist and software engineer with 13 years of experience, currently building scalable systems at Google Switzerland. He brings a strong ML and GNN focus, having contributed key performance and feature improvements to TensorFlow GNN—most notably implementing and hardening a MultiHeadAttentionConv layer with trainable score scaling and thorough tests. Based in Zurich, he blends academic rigor from Durham University's Department of Computer Science with production-grade engineering practices. Known for digging into attention mechanisms and TensorFlow internals, he combines research-minded problem solving with pragmatic delivery at scale.
13 years of coding experience
German, English, French, Spanish
Stackoverflow
1,354reputation
79kreached
35answers
3questions
cuda
top-5%
Github Skills (17)
python10
gnn10
machine-learning10
deeplearning-ai10
deep-learning10
tensorflow10
keras9
cuda9
architecture7
architectures7
nvidia6
sse6
gpu6
struct6
debugging6
Programming languages (6)
Github contributions (5)
tensorflow/gnn
Sep 2022 - Jan 2023
TensorFlow GNN is a library to build Graph Neural Networks on the TensorFlow platform.
Role in this project:
ML Engineer
ML EngineerContributions:7 commits in 3 months
Contributions summary:Pedro primarily contributed to the TensorFlow GNN library by implementing and refining the `MultiHeadAttentionConv` layer. Their work included optimizing performance, introducing features like trainable score scaling, and ensuring correct application of attention activations. The user also added tests to validate the functionality and different configurations of the `MultiHeadAttentionConv` layer. These changes demonstrate a strong understanding of GNN architectures and TensorFlow.
gnndeep-learningneural-graphneural-networksmachine-learning
gonnet/tensorflow
Jul 2018 - Nov 2020
Computation using data flow graphs for scalable machine learning
Contributions:4 pushes in 2 years 4 months
computationscalabledata-sciencemachine-learninggraphs
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.