Jiyan Yang
Senior Tech Lead Manager at Meta
Palo Alto, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJiyan Yang is a Senior Tech Lead Manager at Meta with a PhD in Computational Mathematics from Stanford and 11 years of experience building scalable machine learning and big-data systems. He blends deep theory in randomized linear algebra and large-scale optimization with hands-on engineering—contributing to projects like the Caffe2 deep learning framework where he implemented layers, batch normalization, and optimization improvements. At Meta he progressed from Research Scientist to leading applied AI research efforts that translate academic methods into production platforms on Hadoop and Spark. His career spans collaborations with national labs, finance, and industry research labs, reflecting a knack for cross-domain problem solving. Based in Palo Alto, he is especially interested in leveraging ML methods for internet, life-science, and energy applications and in turning provable algorithms into robust, production-ready software.
11 years of coding experience
9 years of employment as a software developer
Bachelor of Science (B.S.), Mathematics, Bachelor of Science (B.S.), Mathematics at Nanjing UniversityDoctor of Philosophy (Ph.D.), Computational Mathematics, Doctor of Philosophy (Ph.D.), Computational Mathematics at Stanford University
Chinese, Chinese, English
Github Skills (11)
machine-learning10
eval10
caffe10
deep-learning10
trainings10
python10
evaluation10
modeling10
ai9
pytorch4
tensorflow4
Programming languages (3)
Github contributions (5)
facebookarchive/caffe2
Mar 2017 - Mar 2018
Caffe2 is a lightweight, modular, and scalable deep learning framework.
Role in this project:
ML Engineer
ML EngineerContributions:54 commits, 4 PRs, 2 pushes in 1 year
Contributions summary:Jiyan primarily contributed to the Caffe2 deep learning framework by implementing and modifying layers and workflows related to model training, evaluation, and prediction. They developed a new "MapToRange" layer for handling raw ID inputs and indexing, and added and modified several loss layers, including changes for evaluation and removal of unnecessary operations from the eval net. Furthermore, the user worked on several other optimization related changes and enhancements, including the addition of batch normalization and gradient clipping.
pytorchscalablecaffe2deep-learningml
chocjy/pytorch
Apr 2018 - Jul 2020
Tensors and Dynamic neural networks in Python with strong GPU acceleration
Contributions:53 pushes, 21 branches in 2 years 4 months
pythongpu-accelerationdeep-learninggpuacceleration
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.