Orion Reblitz-richardson
AI Researcher And Consultant at Distiller Labs
Honolulu, Hawaii, United States
Summary
🤩
Rockstar🎓
Top SchoolOrion Reblitz-Richardson is an AI researcher and consultant based in Honolulu with a decade of engineering and leadership experience spanning deep learning frameworks, interpretability, and on-device AI. Formerly a senior engineering leader at Meta and contributor at PyTorch and the Meta Superintelligence Lab, he helped launch ExecuTorch, Captum, and PyTorch Foundation initiatives that bridged research, tooling, and production deployment. His work uniquely blends software/hardware co-design, developer ergonomics, and visualization—delivering perf infra, profilers, and privacy-aware on-device training for constrained hardware. Now independent, he focuses on alignment, ethics, morality, and corpus-level pretraining decisions while doing selective consulting from Hawaii. Colleagues know him for refactoring large open-source codebases (TensorBoard, Captum, Caffe2) to improve maintainability and portability, and for turning research ideas into shipped engineering products.
10 years of coding experience
21 years of employment as a software developer
Master of Engineering - MEng, Electrical Engineering & Computer Science, Master of Engineering - MEng, Electrical Engineering & Computer Science at Massachusetts Institute of Technology
Stackoverflow
101reputation
11kreached
2answers
0questions
Github Skills (38)
pytorch10
caffe10
interpretation10
c-language10
python10
protocol-buffers10
testing10
machine-learning10
build-system10
cmake10
tensorboard10
deeplearning-ai10
deep-learning10
cprogramming-language10
protobuf10
Programming languages (10)
Github contributions (5)
facebookarchive/caffe2
Dec 2016 - Sep 2018
Caffe2 is a lightweight, modular, and scalable deep learning framework.
Role in this project:
Back-end Developer & DevOps Engineer
Back-end Developer & DevOps EngineerContributions:118 commits, 139 PRs, 47 pushes in 1 year 9 months
Contributions summary:Orion primarily contributed to the build and documentation aspects of the Caffe2 deep learning framework. Their commits involved modifications to build scripts for different operating systems (macOS), including dependency updates, and addressing documentation generation issues using CMake and Python scripts. Additionally, they were responsible for integrating and then reverting changes to the Jenkins docker base, suggesting a DevOps focus on CI/CD infrastructure. Further contributions include potential fixes to core net testing failures and cleanup of operator documentation for catalog generation.
pytorchscalablecaffe2deep-learningml
pytorch/captum
Aug 2019 - Dec 2019
Model interpretability and understanding for PyTorch
Role in this project:ML Engineer
ML EngineerContributions:72 commits, 24 PRs, 9 pushes in 3 months
Contributions summary:Orion primarily worked on refactoring and reorganizing the `captum` library's attribution methods, moving code into internal or core modules. They updated import statements and restructured file paths, reflecting a focus on code organization and maintainability. Their contributions involved changes across various attribution algorithms (IntegratedGradients, DeepLift, Saliency, etc.) and related testing files, demonstrating familiarity with the core functionality of the library. Additionally, the user contributed to the setup and configuration of the Captum Insights build process, implying a broader involvement in the project's development lifecycle.
pytorchinterpretable-aifeature-importanceunderstandinginterpretability
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.