Scholar
Daiqing Li
Google Scholar ID: 8q2ISMIAAAAJ
University of Toronto, Roblox, ex Playground, NVIDIA Research
Citations & Impact
Citations
1,905
H-index
13
i10-index
14
Publications
20
Co-authors
16
list available
Publications
Resume (English only)
Academic Achievements
- Mar 2025: Published position paper 'Cube: A Roblox View of 3D Intelligence' on arXiv and open-sourced Cube3D, a text-to-3D foundation model
- Oct 2024: Published 'Playground v3', introducing a novel architecture that deeply fuses Llama3 to achieve SOTA in text generation and text-image alignment
- Feb 2024: Open-sourced 'Playground v2.5', achieving better aesthetic quality than Midjourney 5.2
- Jul 2023: Paper 'DreamTeacher' accepted to ICCV 2023, proposing a pretraining framework that distills knowledge from generative models to image backbones
- 2023: Co-authored 'NeuralField-LDM', published at CVPR 2023
- Dec 2023: Gave a talk at University of Bern
- Recipient of MICCAI Young Scientist Award (Runner-up)
Research Experience
- Currently Principal AI Scientist at Roblox, working on 3D foundation models
- Former Research Lead at Playground, leading the development and open-sourcing of state-of-the-art image generative models: Playground-v2, v2.5, and v3
- Former Senior Research Scientist at NVIDIA Toronto AI Lab, working on computer vision, graphics, generative models, and machine learning
- Collaborated closely with Sanja Fidler and Antonio Torralba at NVIDIA
- Multiple research outcomes integrated into NVIDIA products such as Omniverse and Clara
