Xingqian Xu
Senior Research Scientist at NVIDIA
San Francisco Bay Area United States
Summary
🤩
Rockstar🎓
Top SchoolXingqian Xu is a Senior Research Scientist based in the San Francisco Bay Area with a decade of experience specializing in Generative AI across text-to-image, text-to-3D, and text-to-video systems. A UIUC Ph.D., he is first author on influential works like Versatile Diffusion (ICCV23) and several CVPR papers that advance diffusion models, editing, and latent space smoothness. At Picsart and MeshyAI he led product-focused research to build and deploy photorealistic text-to-image models and explore 3D generative pipelines, and he currently advances GenAI at NVIDIA. His engineering work includes practical inference optimizations (float16 support and multi-version model handling) for the well-starred Versatile-Diffusion repo, reflecting both research depth and production savvy. He combines rigorous academic publication impact with hands-on model deployment and data-format expertise (voxels, meshes, watertight manifolds), enabling end-to-end GenAI productization.
10 years of coding experience
7 years of employment as a software developer
Shenzhen Middle SchoolDoctor of Philosophy - PhD Computer Vision & Machine Learni, Doctor of Philosophy - PhD Computer Vision & Machine Learni at University of Illinois Urbana-Champaign
Chinese, Chinese, English
Github Skills (10)
continuous-deployment10
diffusion-models10
pytorch10
machine-learning10
python10
ml-deployment10
f9
computer-vision9
api-design7
api7
Programming languages (1)
Github contributions (5)
SHI-Labs/Versatile-Diffusion
Nov 2022 - Feb 2023
Versatile Diffusion: Text, Images and Variations All in One Diffusion Model, arXiv 2022 / ICCV 2023
Role in this project:
ML Engineer
ML EngineerContributions:66 commits, 1 PR, 48 pushes in 3 months
Contributions summary:Xingqian primarily focused on modifying the `inference.py` file within the repository, which suggests a focus on the practical application and deployment of the Versatile Diffusion model. The commits indicate efforts to incorporate and manage different model versions (e.g., "official," "dc") and to support various functionalities. Furthermore, the user introduced support for float16 precision, optimizing model inference.
pytorchdeep-learningall-in-onecomputer-visiongenerative-model
SHI-Labs/Rethinking-Text-Segmentation
Nov 2020 - Sep 2022
[CVPR 2021] Rethinking Text Segmentation: A Novel Dataset and A Text-Specific Refinement Approach
Contributions:27 commits, 11 PRs, 10 pushes in 1 year 9 months
refinementapproachdatasetcvprtext-segmentation
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.