Xianzhi Du
Staff Research Scientist at Apple
Saratoga, California, United States
Summary
🤩
Rockstar🎓
Top SchoolXianzhi Du is a Staff Research Scientist with eight years of experience advancing LLM pre-training, mid-training, mixture-of-experts, and multimodal generative AI, currently driving research and product integration at Apple. He combines deep academic training (PhD, University of Maryland) with hands-on engineering from roles at Google Brain and Samsung, where he helped build TensorFlow’s official computer vision libraries and deploy state-of-the-art models into production. His contributions to high-profile open-source projects include improving Mask R-CNN with SpineNet backbones and enhancing RetinaNet data pipelines, reflecting a practical focus on robust detection and augmentation. Based in Saratoga, California, he bridges foundational research and scalable system design, often working at the intersection of computer vision and large-scale language models. An uncommon thread in his work is translating advanced model architectures into production-ready components that improve both accuracy and deployment efficiency.
8 years of coding experience
11 years of employment as a software developer
Doctor of Philosophy (Ph.D.), Computer Science, Doctor of Philosophy (Ph.D.), Computer Science at University of Maryland
English, Chinese
Github Skills (7)
mask-rcnn10
object-detection10
computer-vision10
machine-learning10
tensorflow10
retinanet9
data-augmentation8
Programming languages (2)
Github contributions (5)
tensorflow/models
Apr 2021 - Jun 2022
Models and examples built with TensorFlow
Role in this project:
ML Engineer
ML EngineerContributions:27 reviews, 78 commits, 22 comments in 1 year 2 months
Contributions summary:Xianzhi contributed to the Mask R-CNN model within the TensorFlow models repository. Their changes primarily involve modifications to the model's architecture, particularly focusing on the integration of SpineNet backbone and the refinement of detection heads. These changes suggest a focus on improving object detection performance and incorporating more advanced model designs within the existing framework. Additionally, the user worked on data loading and augmentation for RetinaNet, including the addition of attribute learning.
deep-learningtensorflow
xianzhidu/axlearn
Jul 2023 - Mar 2024
Contributions:44 pushes, 31 branches in 8 months
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.