Matthias Fey
Software Engineer at Kumo.AI
Dortmund, North Rhine-Westphalia, Germany
Summary
🤩
Rockstar🎓
Top SchoolMatthias Fey is a software engineer and researcher with 11 years of experience, best known as the creator of PyG (PyTorch Geometric) and a founding engineer at Kumo.AI where he leads ML and research efforts. He holds a PhD in computer science from TU Dortmund and combines deep academic expertise in graph representation learning with hands-on systems work—implementing high-performance C++/CUDA extensions for PyTorch (pytorch_cluster, pytorch_scatter, pytorch_sparse). His contributions to widely used open-source projects like PyTorch Geometric and OGB demonstrate both production-grade engineering and research-minded design for scalable GNN workloads. Based in Dortmund, he bridges theory and practice by shipping optimized graph algorithms and tooling that power large-scale graph ML benchmarks and applications.
11 years of coding experience
Bachelor of Science - BS, Informatik, Bachelor of Science - BS, Informatik at Technische Universität Dortmund
Github Skills (25)
algorithms10
pytorch10
c-language10
operation10
python10
gnn10
testing10
tensorrt10
sparse-matrix10
machine-learning10
tensorflow10
autograd10
cuda10
tensor10
rust10
Programming languages (11)
Github contributions (5)
pyg-team/pytorch_geometric
Oct 2017 - Jan 2023
Graph Neural Network Library for PyTorch
Role in this project:
Software Engineer
Software EngineerContributions:43 releases, 2086 reviews, 4945 commits in 5 years 4 months
Contributions summary:Matthias contributed to the PyTorch Geometric library, focusing on improvements and feature additions in graph neural network implementations. They fixed compatibility issues in the Planetoid datasets related to NumPy, added helper functions for module equivalence and invariance testing, and implemented edge weight support to the LightGCN model. Further contributions include updates to the batching tutorial with cleaner code and the addition of a test to verify the correctness of the ClusterLoader.
pytorchgraph-convolutional-networksgeometric-deep-learningdeep-learningneural-graph
rusty1s/pytorch_cluster
Jan 2018 - Jan 2023
PyTorch Extension Library of Optimized Graph Cluster Algorithms
Role in this project:Back-end Developer
Back-end DeveloperContributions:39 releases, 18 reviews, 548 commits in 5 years 1 month
Contributions summary:Matthias appears to be primarily involved in the development of the PyTorch extension library for optimized graph cluster algorithms. The commits demonstrate the creation of core functionality, including the implementation of a grid clustering algorithm, along with related dependencies. Their work focuses on building out core libraries, including creating C++ implementations to take advantage of CUDA.
pytorchgeometric-deep-learningcluster-algorithmsmachine-learninggraph-neural-networks
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.