Yi Zhu
Co-Founder at Boson AI
United States
Summary
🤩
Rockstar🎓
Top SchoolYi Zhu is a founder and deep learning researcher with a decade of experience building AI-driven products and talent systems that scale, currently leading OpenPaideia and co-founding Boson AI where he grew enterprise revenue into eight figures. His background spans applied research and productization at AWS—contributing to multimodal understanding efforts in SageMaker and Prime Video—and earlier research internships at NVIDIA and other labs focused on computer vision and representation learning. Yi is an active open-source contributor to prominent projects like AutoGluon, where he advanced image classification, CLIP-based zero-shot methods, and object detection integrations, reflecting a strength in bridging research models with production tooling. He holds a PhD in Computer Science from UC Merced and blends rigorous academic training with hands-on deployment experience across startups and cloud-scale teams. Notably, Yi designs systems that marry AI and human workflows to create scalable education and talent engines, a throughline connecting his research and entrepreneurial work.
10 years of coding experience
8 years of employment as a software developer
University of KansasDoctor of Philosophy (PhD) Computer Science, Doctor of Philosophy (PhD) Computer Science at University of California, Merced
Chinese, English
Github Skills (16)
object-detection10
computer-vision10
pytorch10
machine-learning10
deeplearning-ai10
deep-learning10
multimodal10
automated-machine-learning10
clip10
python10
image-classification10
scikit-learn6
scikit6
tabular5
datatable5
Programming languages (5)
Github contributions (5)
autogluon/autogluon
Jun 2022 - Dec 2022
Fast and Accurate ML in 3 Lines of Code
Role in this project:
ML Engineer
ML EngineerContributions:348 reviews, 15 commits, 35 PRs in 5 months
Contributions summary:Yi made several contributions to the AutoGluon library, primarily focused on enhancing the image classification and object detection capabilities. They implemented tutorials for image classification, including zero-shot image classification using CLIP, and added CLIP embedding functionality. Additionally, they integrated MMdetection for object detection inference and addressed bug fixes and improvements in the image prediction module, demonstrating expertise in computer vision and model integration within the AutoGluon framework. The user also contributed to the deprecation of existing image prediction and object detection modules, guiding users toward the more advanced multimodal features.
forecastingimage-textmlppythonmeta-learning
bryanyzhu/Hidden-Two-Stream
Jun 2017 - Dec 2017
Contributions:6 commits, 4 pushes, 1 branch in 5 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.