Towaki Takikawa
Co-Founder / CEO
Outerport
I am currently working on complex document understanding, using iterative reasoning and a hierarchical caching system specialized for tensors and KV cache. If this sounds interesting, please reach out to towaki@outerport.com.
Until recently, I worked as a Research Scientist at NVIDIA Research on the Hyperscale Graphics Systems team with David Luebke and Morgan McGuire.
I was also concurrently a Computer Science PhD student (on leave) at the University of Toronto, supervised by Sanja Fidler and Alec Jacobson, supported by the Royal Bank of Canada, Borealis AI Fellowship.
My interests broadly span machine learning, systems, computer graphics, and computer vision. In particular, I'm interested in compression, retrieval, distribution formats, and infrastructure for multimedia. In the past, I have worked on neural fields [1] and their combination with differentiable data structures [2,3,4] as a new compressed data format for immersive media (3D JPEG!)- as well as foundation models [5] for generating them.
Experience
Publications
Eurographics 2025
FastAtlas: Real-Time Compact Atlases for Texture Space Shading
Nicholas Vining, Zander Majercik, Floria Gu, Towaki Takikawa, Ty Trusty, Paul Lalonde, Morgan McGuire, Alla Sheffer
GPU-based atlas packing for real-time texture space rendering.
arXiv 2024
A LoRA is Worth a Thousand Pictures
Chenxi Liu, Towaki Takikawa, Alec Jacobson
LoRAs > embeddings for retrieval and clustering.
CVPR 2024
What You See is What You GAN: Rendering Every Pixel for High-Fidelity Geometry in 3D GANs
Alex Trevithick, Matthew Chan, Towaki Takikawa, Umar Iqbal, Shalini De Mello, Manmohan Chandraker, Ravi Ramamoorthi, Koki Nagano
Adaptive sampling in rendering = higher fidelity geometry.
SIGGRAPH Asia 2023
Compact Neural Graphics Primitives with Learned Hash Probing
Towaki Takikawa, Thomas Müller, Merlin Nimier-David, Alex Evans, Sanja Fidler, Alec Jacobson, Alexander Keller
Learned hash probing + hash tables = more compression & faster.
ICCV 2023
ATT3D: Amortized Text-To-3D Object Synthesis
Jonathan Lorraine, Kevin Xie, Xiaohui Zeng, Chen-Hsuan Lin, Towaki Takikawa, Nicholas Sharp, Tsung-Yi Lin, Ming-Yu Liu, Sanja Fidler, James Lucas
Optimize 3D generation over more prompts -> faster convergence.
CVPR 2023
⭐ Highlight (Top 2%)
Magic3D: High-Resolution Text-to-3D Content Creation
Chen-Hsuan Lin*, Jun Gao*, Luming Tang*, Towaki Takikawa*, Xiaohui Zeng*, Xun Huang, Karsten Kreis, Sanja Fidler, Ming-Yu Liu, Tsung-Yi Lin
Scaling up 3D generation with smarter data structures.
SIGGRAPH North America 2022
Variable Bitrate Neural Fields
Towaki Takikawa, Alex Evans, Jonathan Tremblay, Thomas Müller, Morgan McGuire, Alec Jacobson, Sanja Fidler
Learned indexing = more compressive than hash tables and grids.
EG / CGF 2022
Neural Fields in Visual Computing and Beyond
Yiheng Xie, Towaki Takikawa, Shunsuke Saito, Or Litany, Shiqin Yan, Numair Khan, Federico Tombari, James Tompkin, Vincent Sitzmann, Srinath Sridhar
Survey of representing vector fields with neural networks.
Tech Report 2022
RTMV: A Ray-Traced Multi-View Synthetic Dataset for Novel View Synthesis
Jonathan Tremblay, Moustafa Meshry, Alex Evans, Jan Kautz, Alexander Keller, Sameh Khamis, Charles Loop, Nathan Morrical, Thomas Müller, Koki Nagano, Towaki Takikawa, Stan Birchfield
Really big dataset of synthetic multi-view images.
CVPR 2021
⭐ Oral Presentation (Top 4%)
Neural Geometric Level of Detail: Real-time Rendering with Implicit 3D Shapes
Towaki Takikawa*, Joey Litalien*, Kangxue Yin, Karsten Kreis, Charles Loop, Derek Nowrouzezahrai, Alec Jacobson, Morgan McGuire, Sanja Fidler
Differentiable data structures + neural fields + GPU programming = fast.
Education
University of Toronto
Ph.D. in Computer Science
Advisor: Sanja Fidler and Alec Jacobson
University of Waterloo
B.S. in Computer Science
Portfolio
