Jason Wang
Machine Learning Engineer at APACHE SOFTWARE FOUNDATION
Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolJason Wang is a Machine Learning Engineer with 10 years of experience blending low-level systems engineering (C/C++, CUDA, Linux) and high-level ML development in Python across autonomous driving, perception fusion, and model deployment. He has shipped production features for AR HUD visualization, multi-sensor fusion, and vehicle integration at OEM scale, and has hands-on experience with BEV perception, trajectory prediction, and IMU/HDMap solutions. An active open-source contributor, Jason has strengthened PaddlePaddle tooling, FastDeploy model support, and TVM frontend integration—work that helped bridge model conversion and fast on-device inference. His background in IC design and EDA tooling gives him uncommon depth in performance-sensitive engineering and memory management, reflected in contributions to shared-memory utilities and model compression. Based in Beijing, he combines algorithm research, team leadership, and cross-team technical integration to drive end-to-end ML systems from prototype to deployment.
10 years of coding experience
6 years of employment as a software developer
Bachelor's degree, Integrated Circuit Design, Bachelor's degree, Integrated Circuit Design at 哈尔滨理工大学
Github Skills (24)
continuous-deployment10
c-language10
python10
machine-learning10
onnx10
ml-deployment10
tensorflow10
trainings10
paddlepaddle10
computer-vision10
cprogramming-language10
protobuffer10
protobuf10
modeling10
convert9
Programming languages (6)
Github contributions (5)
PaddlePaddle/X2Paddle
Jan 2019 - Nov 2022
Deep learning model converter for PaddlePaddle. (『飞桨』深度学习模型转换工具)
Role in this project:
Backend Developer & ML Engineer
Backend Developer & ML EngineerContributions:7 releases, 141 reviews, 646 commits in 3 years 10 months
Contributions summary:Jason's commits primarily involve modifications to the `framework_pb2.py` file, which is part of a deep learning model conversion tool. These changes include additions to protobuf definitions and the creation of a demo involving loading and saving a ResNet model's checkpoint. This indicates a focus on adapting or extending the conversion capabilities for various TensorFlow models, with an emphasis on TensorFlow to PaddlePaddle conversion.
pytorchx2paddle-model-zoopaddlepaddledeep-learningonnx
PaddlePaddle/PaddleX
Apr 2020 - Dec 2022
All-in-One Development Tool based on PaddlePaddle
Role in this project:ML Engineer
ML EngineerContributions:4 releases, 63 reviews, 459 commits in 2 years 7 months
Contributions summary:Jason's commits primarily involve the modification of a shared memory management utility, particularly focusing on memory allocation and data copying within the `paddlex/cv/datasets/shared_queue/sharedmemory.py` file. They also modified documentation related to the image processing and model training procedures, while also fixing some install issues, and fixing and updating model and deployment. These changes suggest a focus on core framework infrastructure and model training processes, likely for machine learning tasks. They also did work for model compression.
resnetpaddletensorflowclassificationend-to-end
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.