Deyu Huang
Software Engineer at ByteDance
Singapore, Singapore
Summary
🤩
Rockstar🎓
Top SchoolDeyu Huang is a software engineer with 8 years of experience specializing in ML systems and AI framework integration, currently applying machine learning at ByteDance in Singapore. Previously at Microsoft he focused on ONNX converters and TF-ONNX work, contributing bug fixes, new op and datatype support, and test automation to the widely used tensorflow-onnx project. He blends back-end engineering, model conversion expertise, and CI/test-driven development to keep ML toolchains reliable across frameworks like TensorFlow, TFJS and TFLite. Academically grounded with a Master’s in Computer Science and research experience in cyber/computer security, he brings both practical production-facing skills and a strong attention to correctness in model interoperability.
8 years of coding experience
3 years of employment as a software developer
Master's degree, Computer Science, Master's degree, Computer Science at Beijing University of Posts and TelecommunicationsResearch assistant, Cyber/Computer Security, Research assistant, Cyber/Computer Security at Chinese Academy of SciencesBachelor's degree, Information Security, Top 1%, Bachelor's degree, Information Security, Top 1% at Hunan University of Science and Technology
Github Skills (9)
tensorflow10
onnx10
python10
test-automation10
machine-learning9
deep-learning9
keras9
debug8
tflite8
Programming languages (7)
Github contributions (5)
onnx/tensorflow-onnx
Nov 2021 - Sep 2022
Convert TensorFlow, Keras, Tensorflow.js and Tflite models to ONNX
Role in this project:
Back-end Developer & Test Automation Engineer
Back-end Developer & Test Automation EngineerContributions:3 releases, 103 reviews, 67 commits in 10 months
Contributions summary:Deyu primarily contributed to the `tensorflow-onnx` repository by fixing bugs and improving the conversion process for TensorFlow models. Their work involved resolving issues related to fused batch normalization, opset updates, and TFJS CI failures, indicating a focus on maintaining the compatibility and accuracy of the conversion tools. Additionally, the user added support for new features like UINT32/UINT64 data types and Rint op to enhance the ONNX model conversion capabilities. Their contributions also included refactoring code and writing tests to validate model conversion and ensure the reliability of the library.
tensorflowjsexportdeep-learningonnxkeras-tensorflow
hwangdeyu/hwangdeyu.github.io
May 2018 - Nov 2022
Contributions:32 pushes in 4 years 6 months
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.