Yunlu Li
Software Engineer at Meta
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolYunlu Li is a software engineer with 11 years of experience building large-scale systems and ML infrastructure, currently focused on GenAI at Meta in the San Francisco Bay Area. Previously at Google, Yunlu worked on TensorFlow, TensorFlow Lite, and model optimization to enable performant inference, and earlier contributed to Google Search indexing and Oracle storage replication—demonstrating fluency across distributed systems and ML deployment. An active contributor to TensorFlow Model Optimization, Yunlu improved pruning and sparsity features and produced examples combining sparsity with quantization to make models more efficient in production. With advanced training from Carnegie Mellon and additional studies at Stanford and Peking University, Yunlu blends rigorous academic foundations with practical production experience delivering performance-sensitive ML systems.
11 years of coding experience
7 years of employment as a software developer
Bachelor's degree Electrical Engineering and Computer Science, Bachelor's degree Electrical Engineering and Computer Science at Peking UniversityMaster's degree Electrical and Computer Engineering, Master's degree Electrical and Computer Engineering at Carnegie Mellon UniversityComputer Science, Computer Science at Stanford University
Github Skills (11)
keras10
pruning10
machine-learning10
tensorflow10
model-optimization10
model-compression10
python9
deep-learning8
deeplearning-ai8
quantization8
tflite7
Programming languages (2)
Github contributions (5)
tensorflow/model-optimization
May 2019 - Apr 2021
A toolkit to optimize ML models for deployment for Keras and TensorFlow, including quantization and pruning.
Role in this project:
ML Engineer
ML EngineerContributions:5 reviews, 12 commits, 24 comments in 1 year 11 months
Contributions summary:Yunlu primarily contributed to the model optimization toolkit, focusing on improving and expanding sparsity features within the TensorFlow model optimization framework. They implemented and fixed functionalities related to pruning, including saving and restoring pruning information in model checkpoints, and improving the integration of pruning with distribution strategies. Furthermore, the user added new examples demonstrating the combination of sparsity and quantization techniques, and enhanced the existing MNIST training script to incorporate these features. Their work aimed at improving the performance and efficiency of the models through sparsity.
ml-modelspythonquantized-networkstensorflowsparsity
liyunlu0618/mazewar
Apr 2016 - May 2016
Contributions:11 pushes, 1 branch in 1 month
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.