I lead the Creative and Applied AI Tools (CAAT) group within NVIDIA Spatial Intelligence Lab, with the focus on 1) application of AI to creative/interactive tools and research to enable such applications, 2) efforts to accelerate 3D Deep Learning research. Our goal is both to develop new research, and release cutting-edge technology that fuels practical tools for researchers and artists.
News
Nov 2025
We release VoMP paper, vastly outperforming existing solutions for volumetric mechanical property prediction, across representations. Code/models coming soon.
Kudos Rishit Dagli and team!
Oct 2025
Am honored to serve on the a SIGGRAPH Technical Papers committee again in 2026.
Aug 2025
My team showcases robot/splat simulation in the NVIDIA Keynote at SIGGRAPH!
Aug 2025
Kaolin Library hosts a hands-on lab at SIGGRAPH with mixed splat/mesh simulation.
Aug 2025
Gaussian Splat Painting is presented at SIGGRAPH and Real-Time Live!
Kudos Karran Pandey and our collaborators from the National Film Board of Canada!
Aug 2025
Kaolin Library 0.18.0 adds collisions to splat/mesh simulations, Warp speedups and more.
Kudos Clement Fuji Tsang, Vismay Modi and team!
Interactive Demos & Prototypes
NVIDIA Kaolin Library for Interactive 3D AI Research
Publications
Diffusion Texture Painting
SIGGRAPH
2024
In the Press:
Neural Brushstroke Engine: Learning a Latent Style Space of Interactive Drawing Tools
SIGGRAPH Asia (TOG)
2022
3DStyleNet: Creating 3D Shapes with Geometric and Texture Style Variations
ICCV
2021
Nonlinear Color Triads for Approximation, Learning and Direct Manipulation of Color Distributions
SIGGRAPH
2020
Playful Palette: An Interactive Parametric Color Mixer for Artists
SIGGRAPH
2017
Fab Forms: Customizable Objects for Fabrication with Validity and Geometry Caching
SIGGRAPH
2015
Formatting Time-aligned ASR Transcripts for Readability
NAACL HLT
2010
Empathic Painting: Interactive Stylization through Observed Emotional State
NPAR
2006
In the Press:
Doctorate Work
My thesis
The Design of Playful and Intelligent Creative Tools is
here. Digital technology offers an unprecedented ability to craft our own
creative media, and machine intelligence holds boundless potential for
empowering creativity. How does one approach design when nearly anything is possible? This thesis addresses the problem of building new creative tools, with intelligence in the interactive loop, without losing sight of the human creators and their needs. Work collected here has been published as first author paper in the top-tier venues of three major domains, including twice at SIGGRAPH (graphics), CVPR (computer vision), and CHI (human computer interaction).
Go to colorsandbox.com for demos of my work on color.
Talks
