Ori Levari
Software Engineering Manager at Floodbase
San Francisco, California, United States
Summary
🤩
Rockstar🎓
Top SchoolOri Levari is a Software Engineering Manager in San Francisco with a decade of experience leading teams and shipping production-grade systems across startups and Microsoft. He blends hands-on backend and automation engineering with ML integration expertise—contributing to high-profile open-source projects like ONNX Runtime and Windows Machine Learning, including DirectX 12-based model inference samples and build automation for onnxruntime.dll. At Floodbase he progressed from senior engineer to manager, guiding product delivery while retaining deep technical involvement. A University of Washington CS graduate with a minor in Global Health, Ori brings a pragmatic research-to-production mindset informed by early work in humanitarian and health-focused organizations. Colleagues rely on him for clean engineering fixes, build-system resilience, and pragmatic ML deployment patterns that bridge low-level graphics APIs and modern inference runtimes.
9 years of coding experience
7 years of employment as a software developer
Bachelor of Science (B.S.), Computer Science, Bachelor of Science (B.S.), Computer Science at University of Washington
English
Stackoverflow
86reputation
463reached
1answer
0questions
Github Skills (16)
machine-learning10
c-language10
direct3d10
onnx10
cprogramming-language10
build-automation10
directx10
hardware-acceleration9
deep-learning9
cmake8
ai-framework6
python6
windows-machine-learning6
cicd6
computer-vision4
Programming languages (4)
Github contributions (5)
microsoft/Windows-Machine-Learning
Nov 2018 - Mar 2021
Samples and Tools for Windows ML.
Role in this project:
ML Engineer
ML EngineerContributions:1 release, 5 reviews, 51 commits in 2 years 4 months
Contributions summary:Ori primarily contributed to the development of a D3D12 command queue sample, demonstrating expertise in integrating machine learning models with DirectX 12. Their work involved constructing a LearningModelDevice from a D3D12 command queue, loading models, creating sessions and bindings, and running model evaluations. Further contributions included the addition of a custom operator sample, and updating the top3 results code across all samples.
pytorchcaffe2pythonwindowswinmltools
microsoft/onnxruntime
Nov 2019 - Apr 2021
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:Back-end Developer & Automation Engineer
Back-end Developer & Automation EngineerContributions:27 reviews, 49 commits, 99 PRs in 1 year 5 months
Contributions summary:Ori primarily contributed to the maintenance and improvement of the ONNX Runtime by fixing bugs and updating existing code. They addressed issues such as missing IR versions in test data and incorrect ordering within header files. Furthermore, the user was involved in the automation aspects of the project, specifically in the build process for the `onnxruntime.dll` file by populating the file metadata, and in adapting the build process to handle downlevel and WCOS scenarios. The user also implemented fixes related to SDL for better code security and debugging.
runtimetrainingtensorflowai-frameworkaccelerator
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.