daquexian
China
Summary
🤩
RockstarJianhao Zhang is a seasoned programmer with 10 years' experience specializing in ML engineering, model interoperability, and Android build tooling, currently contributing at Skywork AI and serving as an ONNX approver. He has a strong open-source footprint—authoring core logic for the popular onnx-simplifier and improving ONNX model converters and shape inference across projects like Alibaba MNN, OneFlow, and the ONNX repository. Jianhao combines backend systems work with mobile and DevOps strengths, having enabled Android builds and Java APIs for ONNX Runtime and contributed to Thunderbird Android. His track record at OneFlow, JD.com and Microsoft shows practical impact on production ML frameworks, automated testing, and profiling. Based in China and educated at the University of Science and Technology of China, he pairs deep framework-level expertise with a pragmatic focus on portability and build reliability.
10 years of coding experience
Stackoverflow
169reputation
20kreached
5answers
15questions
Github Skills (52)
unit-testing10
pytorch10
ios10
c-language10
python10
sendmail10
package-management10
msn10
setuptools10
send-email10
machine-learning10
deep-q-learning10
inference10
cmake10
onnx10
Programming languages (23)
Github contributions (5)
daquexian/onnx-simplifier
Apr 2019 - Jan 2023
Simplify your onnx model
Role in this project:
Back-end Developer
Back-end DeveloperContributions:26 reviews, 264 commits, 40 PRs in 3 years 9 months
Contributions summary:Daquexian was primarily responsible for developing the core functionalities of the "onnx-simplifier" project. They implemented the Python package structure, including `setup.py`, `__main__.py`, and the simplification logic within `onnx_simplifier.py`. These changes include initial package creation, model simplification logic, and command-line interface setup. They also contributed to the version updates and dependency management.
pytorchdeep-learningonnxonnxruntimesimplify
onnx/onnx
Jul 2018 - Nov 2022
Open standard for machine learning interoperability
Role in this project:ML Engineer
ML EngineerContributions:70 reviews, 32 commits, 51 PRs in 4 years 4 months
Contributions summary:Daquexian primarily contributes to the shape inference capabilities within the ONNX repository, evidenced by commits focused on fixing bugs in shape inference related to convolutional layers and dilated convolutions, as well as other pooling operations. The user implemented tests for these shape inference improvements, ensuring the accuracy of the changes. They also updated the documentation and addressed issues related to the Resize and Softmax operators, demonstrating a focus on maintaining the ONNX standard's functionality.
pytorchmxnetdeep-learninginteroperabilitymachine-learning
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.