Jay Wu
Research Scientist at NVIDIA
Summary
🤩
RockstarJay Wu is a research scientist with six years of hands-on machine learning experience, currently contributing to NVIDIA and based in Singapore. He specializes in diffusion-based generative models and played an instrumental engineering role on the ICCV 2023 Tune-A-Video project, implementing a UNet3D conditioning model and DDIM inversion to enable one-shot text-to-video tuning. Jay blends research rigor with practical tooling—adding training Colabs and core sampling enhancements that make cutting-edge video generation more accessible to practitioners. He is comfortable bridging model architecture, training pipelines, and reproducible demos, demonstrating a knack for translating research ideas into usable code.
6 years of coding experience
Github Skills (9)
transformers10
computer-vision10
pytorch10
machine-learning10
deep-learning10
python10
diffusers10
jupyter-notebook9
notebook9
Programming languages (2)
Github contributions (5)
showlab/Tune-A-Video
Jan 2023 - Mar 2023
[ICCV 2023] Tune-A-Video: One-Shot Tuning of Image Diffusion Models for Text-to-Video Generation
Role in this project:
ML Engineer
ML EngineerContributions:14 commits, 1 PR, 17 pushes in 1 month
Contributions summary:Jay made significant contributions to the `tune-a-video` repository, focusing on core model architecture and training. They implemented the `UNet3DConditionModel`, adapting it from a Hugging Face Diffusers model. They also added DDIM inversion functionality during the sampling process and included a Colab notebook to help users train the model.
showlab/Awesome-Video-Diffusion
Apr 2023 - Mar 2025
A curated list of recent diffusion models for video generation, editing, and various other applications.
Contributions:1 review, 28 PRs, 87 pushes in 1 year 11 months
diffusion-modelsmotion-customizationvideo-editingvideo-generationvideo-generation-evaluation
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.