Bin Luo
Senior Technical Manager at Baidu, Inc.
Changping District, Beijing, China
Summary
🤩
Rockstar🎓
Top SchoolBin Luo is a Senior Technical Manager at Baidu with over a decade of experience building and managing production ML and back-end systems. With a master's in Signal & Information Processing from Beijing Jiaotong University and an early internship at MSRA, he blends strong academic foundations with industry-scale delivery. He contributes to prominent PaddlePaddle projects—PaddleClas and PaddleOCR—implementing MobileNetV1 models and refining advanced loss functions, showing hands-on expertise in model design, training stability, and export for deployment. Bin’s work on pretrained model integration and enhanced CTC variants signals a focus on improving both accuracy and real-world usability of computer vision systems. Based in Changping District, Beijing, he combines leadership at a major AI company with active, practical open-source contributions that help bridge research and production. An understated strength is his pattern of iterating on core algorithmic components, suggesting he favors engineering depth that scales.
10 years of coding experience
Master's degree, Signal & Information Processing, Master's degree, Signal & Information Processing at Beijing Jiaotong University
Stackoverflow
1reputation
0reached
0answers
0questions
Github Skills (17)
python10
pre-trained-model10
image-classification10
model-driven10
ocr10
machine-learning10
model-building10
deep-learning10
mobilenet10
ocra10
paddlepaddle10
computer-vision10
modeling10
model-driven-development10
loss9
Programming languages (6)
Github contributions (5)
PaddlePaddle/PaddleClas
Jan 2021 - Mar 2022
A treasure chest for visual classification and recognition powered by PaddlePaddle
Role in this project:
Back-end Developer & ML Engineer
Back-end Developer & ML EngineerContributions:110 reviews, 249 commits, 147 PRs in 1 year 1 month
Contributions summary:Bin's commits primarily focus on adding and updating a `mobilenet_v1.py` file, which defines a MobileNetV1 model. This suggests involvement in developing or integrating a specific neural network architecture. The updates include code modifications, and the addition of pre-trained weights, indicating the user is likely responsible for model implementation, refinement, and integration with the overall PaddlePaddle classification framework. The frequent updates and inclusion of pretrained models suggest a focus on model accuracy and usability within the PaddlePaddle ecosystem.
classificationpaddlepaddleimage-classificationknowledge-distillationautoaugment
PaddlePaddle/PaddleOCR
Mar 2021 - Feb 2022
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
Role in this project:ML Engineer
ML EngineerContributions:10 reviews, 28 commits, 27 PRs in 11 months
Contributions summary:Bin's primary contribution involves enhancing the enhanced CTC loss function within the PaddleOCR framework. This includes addressing bugs in the Focal-ctc implementation and adding new functionality related to center loss and ace loss. Their work is focused on improving the loss functions used in the OCR model training process and integrating the enhancements with the existing framework, as demonstrated by changes to the loss functions, model export and training script.
aiocrpdfchineseocrpdf2markdown
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.