|
Nathon Lee I’m DeWang Li (nathon-lee), a distributed systems engineer focused on large-scale AI infrastructure and performance optimization. This site documents my work, experiments, and learnings in high-performance systems, LLM training, and real-world engineering problems. I work on system-level optimizations across storage, latency-critical services, and model training pipelines, and contribute to open-source projects such as DeepSpeed. |
|
ResearchI'm interested in computer vision, deep learning, generative AI, and image processing. Most of my research is about inferring the physical world (shape, motion, color, light, etc) from images, usually with radiance fields. Some papers are highlighted. |
|
Radiance Meshes for Volumetric Reconstruction
Alexander Mai, Trevor Hedstrom, George Kopanas, Janne Kontkanen, Falko Kuester, Jonathan T. Barron arXiv, 2025 project page / arXiv Parameterizing a scene with a Delaunay tetrahedralization and a neural field yields a scene representation that is accurate, fast to render, easy to edit, and backwards-compatible. |
|
NeRF-Casting: Improved View-Dependent Appearance with Consistent Reflections
Dor Verbin, Pratul Srinivasan, Peter Hedman, Benjamin Attal, Ben Mildenhall, Richard Szeliski, Jonathan T. Barron SIGGRAPH Asia, 2024 project page / arXiv Carefully casting reflection rays lets us synthesize photorealistic specularities in real-world scenes. |
|
Flash Cache: Reducing Bias in Radiance Cache Based Inverse Rendering
Benjamin Attal, Dor Verbin, Ben Mildenhall, Peter Hedman, Jonathan T. Barron, Matthew O'Toole, Pratul P. Srinivasan ECCV, 2024 (Oral Presentation) project page / arXiv A more physically-accurate inverse rendering system based on radiance caching for recovering geometry, materials, and lighting from RGB images of an object or scene. |
Miscellanea |
|
Design and source code from Jon Barron's website |