Muhammed Balin
Applied Scientist II at Deep Graph Library
California, United States
Summary
🤩
Rockstar🎓
Top SchoolMuhammed Balin is an Applied Scientist II and Ph.D. candidate specializing in High-Performance Computing and Machine Learning, with nine years of experience building scalable GNN and LLM systems. Based in California, he blends academic research from Georgia Tech with industry impact at AWS and NVIDIA, where he optimized LLM quantization and implemented CUDA-accelerated GNN sampling. An active contributor and maintainer on the widely used Deep Graph Library (DGL), he led multi-GPU GraphBolt dataloader development and implemented cooperative minibatching and temporal sampling optimizations for large-scale graphs. His work sits at the intersection of distributed systems, GPU performance, and machine learning, making production-grade research reproducible and fast. Notably, he has extended distributed programming frameworks and shipping high-performance primitives that directly accelerate training on real-world, large-scale graphs.
9 years of coding experience
6 years of employment as a software developer
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at Georgia Institute of TechnologyIstanbul High SchoolNon-degree Student Exchange Program Computer Science, Non-degree Student Exchange Program Computer Science at Columbia UniversityBachelor of Science (BS) Computer Engineering and Mathematics, Bachelor of Science (BS) Computer Engineering and Mathematics at Boğaziçi University
English, German, Turkish
Github Skills (16)
cuda10
data-handling10
data-loading10
pytorch10
graph-neural-network10
parallel-computing10
dataloader10
performance-optimization10
python10
load-data10
gnn10
data-processing10
machine-learning9
c-language8
cprogramming-language8
Programming languages (9)
Github contributions (5)
dmlc/dgl
Nov 2022 - Dec 2022
Python package built to ease deep learning on graph, on top of existing DL frameworks.
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:1 release, 1192 reviews, 4 commits in 29 days
Contributions summary:Muhammed made significant contributions to the DGL library, specifically focusing on performance improvements and additions to the Labor sampling method, which is designed to optimize graph neural networks (GNNs) for large-scale graphs. Their work involved implementing optimizations for CUDA-based graph computations and implementing enhancements to the temporal sampling methods. Furthermore, the user implemented a cooperative minibatching framework, adding features for managing node and edge data to the GraphBolt data loader.
pytorchpythondeep-learningmachine-learninggraph-neural-networks
mfbalin/Concrete-Autoencoders
Jan 2019 - May 2022
Contributions:3 reviews, 16 commits, 1 PR in 3 years 4 months
supervised-learningunsupervised-learningmachine-learningicml-2019feature-selection
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.