Mingxing Tan
Research Scientist TLM at Waymo
Mountain View, California, United States
Summary
🤩
Rockstar🎓
Top SchoolMingxing Tan is a research scientist and TLM with eight years of experience building and optimizing machine learning models and compilers at Google and Waymo, based in Mountain View. He has deep hands-on expertise in object detection and efficient model design, contributing to widely used projects like EfficientDet in TensorFlow and Google AutoML and improving TPU and edge deployment workflows. Mingxing also works on low-level ML infrastructure, enhancing the XLA compiler’s GPU codepaths and fusion/reduction logic to boost runtime efficiency. His background in computer engineering from Cornell and computer science from Peking University complements a profile that bridges research, productionization, and systems-level optimization. Notably, his open-source contributions touch both high-level model innovations and backend compiler improvements—an uncommon combination that helps move models from prototype to performant deployment.
8 years of coding experience
Computer Science, Computer Science at Peking UniversityComputer Engineering, Computer Engineering at Cornell University
Github Skills (24)
c-language10
python10
gpu-programming10
machine-learning10
deep-learning10
tensorflow10
xla10
object-detection10
computer-vision10
compiler10
cprogramming-language10
efficientnet10
image-processing9
operation9
convolutional-neural-networks9
Programming languages (3)
Github contributions (5)
google/automl
Mar 2020 - Aug 2022
Google Brain AutoML
Role in this project:
ML Engineer
ML EngineerContributions:116 reviews, 415 commits, 261 PRs in 2 years 5 months
Contributions summary:Mingxing primarily contributed to the implementation and release of EfficientDet, as evidenced by the commit messages and code changes. They worked on the anchor definition, box decoding, and non-maximum suppression, indicating expertise in object detection algorithms. The user also focused on model post-processing, making changes related to detections, leading to the inclusion of a new feature for mixed precision for inference.
brainefficientnetv2pythonefficientnetdeep-learning
tensorflow/tpu
May 2019 - Mar 2021
Reference models and tools for Cloud TPUs.
Role in this project:ML Engineer
ML EngineerContributions:55 commits, 12 PRs, 30 pushes in 1 year 10 months
Contributions summary:Mingxing primarily contributes to the codebase by updating and improving the AutoAugment utilities and working on the EfficientNet models. Their commits focus on incorporating new policies, and modifying the model architecture. The changes also include modifications related to model export and the inclusion of new functionalities for edge devices.
cloud
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.