Jimmy Yao
Educator
Berkeley, California, United States
Summary
🤩
Rockstar🎓
Top SchoolJimmy Yao is an educator and computational mathematician with a Ph.D. from UC Berkeley and eight years of experience applying deep mathematics to practical ML systems. Based in Berkeley, he bridges rigorous theoretical background with hands-on MLOps and backend engineering, contributing to large-scale distributed training projects like Alpa and the Ray AI compute engine. His open-source work focuses on resource-aware parallelism, cluster integration, and model format interoperability—fixing tricky issues from placement groups to CoreML conversion quirks. At Numerade he translates complex STEM concepts into teachable material while retaining active engineering chops. Colleagues know him for hunting down subtle bugs in distributed workflows and for making deployment-oriented improvements that improve scalability and reproducibility.
8 years of coding experience
Doctor of Philosophy - PhD, Mathematics, Doctor of Philosophy - PhD, Mathematics at University of California, Berkeley
Github Skills (23)
distributed-training10
python10
machine-learning10
coreml10
distributed-computing10
deeplearning-ai10
deep-learning10
tensorflow10
ray10
model-conversion10
data-science9
jax9
data-engineering8
compiler-compiler7
compiler7
Programming languages (5)
Github contributions (5)
microsoft/MMdnn
Apr 2018 - Jun 2019
MMdnn is a set of tools to help users inter-operate among different deep learning frameworks. E.g. model conversion and visualization. Convert models between Caffe, Keras, MXNet, Tensorflow, CNTK, PyTorch Onnx and CoreML.
Role in this project:
Back-end Developer
Back-end DeveloperContributions:140 commits, 40 PRs, 12 pushes in 1 year 1 month
Contributions summary:Jimmy primarily worked on coreml_graph and coreml_parser.py files, modifying and updating the graph-related functionalities. The contributions appear to involve the conversion of CoreML models, potentially including adding support for new layers or operators within the model conversion process. The user's modifications also address issues related to depthwise convolutions and other layers as well as handling issues related to reshape and add layer.
caffe2intertensorflowmodel-conversionoperate
ray-project/ray
May 2022 - Aug 2022
Ray is an AI compute engine. Ray consists of a core distributed runtime and a set of AI Libraries for accelerating ML workloads.
Role in this project:ML Engineer
ML EngineerContributions:78 reviews, 16 commits, 54 PRs in 2 months
Contributions summary:Jimmy contributed to the Ray project by addressing issues in the Ray Datasets and AIR components, specifically fixing label tensor squeezing and the type infer of pandas dataframes. They also refactored the ScalingConfig key validation in the AIR module. Additionally, the user was involved in implementing an end-to-end TensorFlow example and setting the correct GPU ID in the TorchTrainer, and making minor adjustments to documentation files.
pythonconsistsruntimetensorflowserving
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.