Quei-an Chen
AI Research Scientist at Luma AI
Taipei, Japan
Summary
🤩
Rockstar🎓
Top SchoolBarkley Dai is a product leader and ML practitioner with 10 years of experience building generative and computer-vision products at the intersection of AI, AR, and creator platforms. As a founding product lead at Luma AI he defines model capabilities and curates datasets powering features like multi-reflection, visual reasoning, and layering, and previously drove rapid user growth and creative tooling at TikTok where he shipped multiple GenAI effects and avatar initiatives. He pairs hands-on ML engineering—contributing to NeRF and instant-ngp PyTorch implementations and CUDA-accelerated pipelines—with product strategy, moving research prototypes into scalable creator-facing features. Based in Taipei and fluent in cross-cultural product expansion, he’s launched viral experiences (0 to 1M users in five days) and advised virtual-influencer ecosystems. A curious maker with a creative streak, he prototypes with Claude Code, runs a VTuber channel on deep-learning topics, and experiments with visualized codebases and short-personalized dramas to explore human creativity and spirituality.
10 years of coding experience
5 years of employment as a software developer
Master of Science - MS, Computer Science, Master of Science - MS, Computer Science at University of OxfordMaster's degree, Master's degree at École Polytechnique
Chinese, Japanese, French, English
Github Skills (13)
3d-reconstruction10
algorithm10
cuda10
pytorch10
serf10
optimisation10
optimizers10
optimization10
pytorch-lightning9
python8
computer-graphics8
linear-algebra7
deep-learning7
Programming languages (9)
Github contributions (5)
kwea123/nerf_pl
Apr 2020 - Oct 2022
NeRF (Neural Radiance Fields) and NeRF in the Wild using pytorch-lightning
Role in this project:
ML Engineer
ML EngineerContributions:10 releases, 5 reviews, 337 commits in 2 years 6 months
Contributions summary:Quei-an's commits primarily focus on modifying and improving the `NeRF` (Neural Radiance Fields) model, specifically within the context of the `pytorch-lightning` framework. The user implemented various optimization algorithms, including RAdam, PlainRAdam, and AdamW, to improve training performance. Furthermore, changes indicate modifications to the rendering process and dataset handling, which are essential for the NeRF model's functionality.
ray-tracingcolabpytorch-lightningnerf-wradiance
kwea123/ngp_pl
Aug 2022 - Dec 2022
Instant-ngp in pytorch+cuda trained with pytorch-lightning (high quality with high speed, with only few lines of legible code)
Role in this project:ML Engineer
ML EngineerContributions:2 releases, 41 commits, 4 PRs in 4 months
Contributions summary:Quei-an primarily contributed to the implementation and improvement of a neural radiance field (NeRF) model, as indicated by the repository description and commit messages. Their work involved modifying code related to volume rendering, including adding sample point weights and integrating a distortion loss function, suggesting an effort to optimize the model's performance and accuracy. The commits also included adapting image reading for the Colmap dataset and adding mesh extraction functionalities, indicating the user's involvement in the broader pipeline related to 3D reconstruction and novel view synthesis.
cudapytorchspeedlineslightning
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.