Cheng Wang

Cheng Wang

About

Publications

Projects

Experience

Cheng Wang

王呈

Builder

Email Github LinkedIn Resume

About Me

Hi! I'm Cheng Wang, a Computer Vision Research Engineer with a strong passion for Visual Computing, particularly inverse problems and physically-based rendering. I earned my Master's degree in Computer Science from UC San Diego, advised by Prof. Tzu-Mao Li, and my Bachelor's degree in Computer Science and Technology at Chu Kochen College of Zhejiang University, advised by Prof. Rui Wang.

My research has been published at CVPR 2025, CVPR 2024, and ACM SIGGRAPH 2026. I have industry experience at Dandy, Electronic Arts, and Rays Engine. Feel free to take a look at my Resume.

Skills

Python

C/C++

PyTorch

OpenCV

Mitsuba

OpenGL

Unreal Engine 5

Blender

Docker

TypeScript

React

Interests

Badminton

Video Games

Photography

Poker

Classical Music

Cooking

Experience

Apr 2025 - June 2026

Computer Vision Research Engineer

Dandy, Remote, United States

Aug 2024 - Apr 2025

Research Assistant

UC San Diego, CA, United States

June 2023 - Sep 2023

Rendering Intern

Electronic Arts, Remote, United States

Sep 2022 - June 2024

M.S. in Computer Science

UC San Diego, CA, United States

Dec 2021 - Jun 2022

Rendering Engine Engineer Intern

RaysEngine, Zhejiang, China

Sep 2018 - Jun 2022

B.E. in Computer Science and Technology

Zhejiang University, Zhejiang, China

Publications and Preprints

A Generalizable Light Transport 3D Embedding for Global Illumination

Bing Xu, Mukund Varma T, Cheng Wang, Tzumao Li, Lifan Wu, Bartlomiej Wronski, Ravi Ramamoorthi, Marco Salvi

ACM SIGGRAPH 2026

Global light transport modeling on millions of triangles via a scalable 3D embedding with point clouds and linear-complexity transformers.

Paper Code Coming Soon!

HotSpot: Screened Poisson Equation for Signed Distance Function Optimization

Zimo Wang*, Cheng Wang*, Taiki Yoshino, Sirui Tao, Ziyang Fu, Tzu-Mao Li

CVPR 2025 (Highlight)

Neural Signed Distance Field optimization from point clouds with heat method.

Paper Project Page Code

Near-Field Lighting Estimation via Ray Regression

Cheng Wang, Tzu-mao Li

Master Thesis

Better near-field lighting estimation by overparameterizing lighting locations as ray bundles.

ZeroRF: Zero-shot Sparse View 360° Reconstruction

Ruoxi Shi*, Xinyue Wei*, Cheng Wang, Hao Su

CVPR 2024

A sparse view 360° Neural Radiance Field reconstruction method inspired by Deep Image Prior.

Paper Project Page Code

Projects

Volumetric Path Tracer

A volumetric path tracer that can handle multiple chromatic heterogeneous volumes with absorption and multiple-scattering, with both phase function sampling and next event estimation, based on an educational physically-based renderer, lajolla.

Autonomous Vehicle Motion Forecasting

A course project using deep learning models to forecast autonomous vehicle motion, predicting object positions three seconds ahead.

Implementation of Tensorial Radiance Field (TensoRF) with PyTorch

TensoRF represents a radiance field by matrix-vector and vector-vector combinations of tensors. Compared to MLP based NeRF, it is more efficient while preserving comparable quality without customized CUDA implementation.

Real-Time Texture-Space Subsurface Scattering

An efficient method to compute convolution of radiance map and weight kernel in texture space. By pre-calculating weight kernels with Burley's normalized diffusion profile and applying wavelet transformation, the method significantly reduced time complexity of convolution.

Awarded the Outstanding Graduation Thesis of Zhejiang University Undergraduates in 2022.

Tiled Forward Shading

An implementation of tiled forward rendering and tiled deferred rendering, with optimized light culling pass by using multiple frustum dividing and light-frustum intersection strategies.

Cyber Creed

2021

A 3D computer game resembling Assassin's Creed with Unreal Engine 4.

Medical Appointment System

A course project implementing a hospital website with Vue.js and Java, whose main functions including information viewing, registration appointment, online consultation, video diagnosis, online forum, etc.

Petiu: Pet farewell service system

An IOS App "Petiu" written in SwiftUI.

Awarded Third Prize in the China Collegiate Computing Contest -- Mobile Application Innovation Contest held by Apple Inc. and Zhejiang University in 2021.

© Cheng Wang. All Rights Reserved.