Shuoxin Lin
Senior Machine Learning Engineer at Axon
Old Toronto, Ontario, Canada
Summary
🤩
Rockstar🎓
Top SchoolShuoxin Lin is a Senior Machine Learning Engineer with eight years of industrial R&D experience building on-device ML systems, MLOps platforms, and privacy-preserving learning at Axon and formerly Apple Core ML and Camera. He holds a PhD in Computer Engineering from the University of Maryland and combines deep research chops with production engineering—authoring Core ML converters (tf-coreml, coremltools) and model compression/low-precision runtimes used for mobile vision and NLP. Shuoxin is a polyglot coder (Python, C++, CUDA, Java, Matlab) who has repeatedly accelerated systems via custom scheduling and GPU/CUDA optimizations, including a published 20x throughput speedup in OFDM transceivers. He also brings interdisciplinary depth from a biomedical engineering undergraduate background and hands-on embedded/hardware experience teaching FPGA labs and building sensor systems, which informs his practical approach to edge ML.
8 years of coding experience
10 years of employment as a software developer
High School, Physics, High School, Physics at South China Normal University High SchoolThe University of Maryland, College ParkDoctor of Philosophy (Ph.D.), Computer Engineering, 4.0, Doctor of Philosophy (Ph.D.), Computer Engineering, 4.0 at University of MarylandBachelor of Science (BS), Biomedical/Medical Engineering, 3.9/4.0, Bachelor of Science (BS), Biomedical/Medical Engineering, 3.9/4.0 at Zhejiang University
Chinese, English
Github Skills (7)
keras10
machine-learning10
model-conversion10
tensorflow10
python10
coreml10
pytorch9
Programming languages (1)
Github contributions (5)
tf-coreml/tf-coreml
Oct 2017 - Jun 2019
TensorFlow to CoreML Converter
Role in this project:
ML Engineer
ML EngineerContributions:89 commits, 9 PRs, 9 pushes in 1 year 8 months
Contributions summary:Shuoxin's initial commit sets up the foundation of the project, introducing the core functionality for converting TensorFlow models to CoreML. They then added support for one-hot encoding, a crucial step for preparing categorical data for machine learning models. The subsequent commits involve fixing bugs, optimizing the conversion process, and adding features to support more TensorFlow operations, demonstrating a deep understanding of model conversion and CoreML integration.
coremltensorflowconverter
apple/coremltools
Aug 2017 - Oct 2019
Core ML tools contain supporting tools for Core ML model conversion, editing, and validation.
Role in this project:ML Engineer
ML EngineerContributions:64 commits, 36 PRs, 14 pushes in 2 years 2 months
Contributions summary:Shuoxin primarily contributed to the Core ML Tools repository, focusing on model conversion and support for various machine learning frameworks. Their work involved implementing separable convolution layers and updating existing code related to Keras model conversion, indicating expertise in translating Keras models to the Core ML format. Further contributions show efforts in supporting different Keras features and configurations, making the conversion process more robust.
pytorchmodel-conversionvalidationdeep-learningmachine-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.