Chunjiang Liu

CMU ECE M.S.

3D Vision Stealth AI Startup AI Agents

chunjiang.liu.richard (at) gmail.com

About Me

I received my M.S. in ECE from Carnegie Mellon University, where I worked with Prof. László A. Jeni and Yizhou Zhao.

Research Interests

AI Agents and World Models
3D Vision and Physical World Modeling

Publications

ICLR

MOSIV: Multi-Object System Identification from Videos

Chunjiang Liu, Xiaoyuan Wang, Qingran Lin, Albert Xiao, Haoyu Chen, Shizheng Wen, Hao Zhang, Lu Qi, Ming-Hsuan Yang, Laszlo A. Jeni, Min Xu, Yizhou Zhao

International Conference on Learning Representations (ICLR), 2026.

PDF

ICCV

Toward Material-Agnostic System Identification from Videos

Yizhou Zhao, Haoyu Chen, Chunjiang Liu, Zhenyang Li, Charles Herrmann, Junhwa Hur, Yinxiao Li, Ming-Hsuan Yang, Bhiksha Raj, Min Xu

IEEE/CVF International Conference on Computer Vision (ICCV), 2025.

PDF Code

ICCV-W

Total-Editing: Head Avatar with Editable Appearance, Motion, and Lighting

Yizhou Zhao, Chunjiang Liu, Haoyu Chen, Bhiksha Raj, Min Xu, Tadas Baltrusaitis, Mitch Rundle, HsiangTao Wu, Kamran Ghasedi

IEEE/CVF International Conference on Computer Vision Workshop (ICCV Workshop), 2025.

PDF

Working Experience

Member of Technical Staff, Bake AI Inc.

Jan. 2026 — Jun. 2026

Built 1Bench, a living benchmark platform tracking frontier AI on expert-verified, unsolvable problems across multiple domains.
Built ArtArena, a visual aesthetic benchmark comparing AI models on artist-curated artworks and expert judgments.

Research Associate, Carnegie Mellon University, ECE

2024 — 2025

Worked on multi-object dynamic reconstruction and physics estimation from video through differentiable simulation.
Developed material-agnostic dynamic reconstruction with learnable neural constitutive models.
Worked on joint control of appearance, motion, and lighting for head avatar editing.

Education

Carnegie Mellon University

M.S. in Electrical and Computer Engineering, 2023 — 2024
GPA: 3.83 / 4.00

University of Electronic Science and Technology of China

B.E. in Communication Engineering, 2018 — 2022
GPA: 3.84 / 4.00

Academic Projects

Fall 2024 · Advisor: Daniel Fried

Weighted angular-distance loss for navigation decisions.
LLM-based instruction paraphrasing for data augmentation.
53.2% success rate on unseen environments in R2R.

Zero-Shot Text-to-Video Generation

Spring 2024 · Advisor: Giulia Fanti

LLM-based prompt elaboration and few-shot frame consistency.
CLIP score: 30.4, compared with 29.6 for the baseline.

Lightweight Transformer for Next-Word Prediction

Spring 2024 · Advisor: Yuejie Chi

Knowledge distillation on DailyDialog.
Perplexity of 35, comparable to GPT-2.

Powered by Jekyll and Minimal Light theme.