Jie Deng is a research assistant at Zhejiang University, advised by Prof. Gaoang Wang (CVNext Lab), He obtained his B.S. in Electrcal Engineering from University of Washington, Seattle, and his M.S. in Electrical and Computer Engineering from Georgia Institute of Technology. Jie stays curious about the wider landscape of computer vision and deep learning, and actively seeks new collaboration opportunities. His work centers around virtual environment simulation/generation, generative models, and LLM reasoning. He is a highly self-motivated, and curious student applying to Ph.D. programs for 2026 Fall. He is also actively looking for research opprtunities. (Please check out my CV for more details.)

My general research direction is AI + City. I aim to develop, leverage, and improve computer vision, large language models (LLMs), and deep learning so they work robustly in complex urban environments. The long-term goal is a high-fidelity virtual city that can simulate everything.

View full research details and examples

• I. Urban Scene Simulation / Generation — Combining generative models, procedural generation, and engines. Details & examples
• II. Urban Scene Reconstruction — NeRF, Gaussian Splatting for high-fidelity city-scale assets. Details & examples
• III. Urban Scene Understanding — 3D understanding, mapping, semantics, LLM-assisted planning. Details & examples
• IV. Traffic and Crowd Simulation — Data-driven and agent-based mobility modeling. Details & examples
• V. Urban Data Analytics — Learning from heterogeneous sensing for forecasting/decision making. Details & examples

Going forward, I plan to continue exploring these directions while keeping up with SOTA progress in DL, CV, and NLP, and adapting those methods to urban problems in principled, task-appropriate ways.

Professional Service