Graphics developer / AI researcher
buildAI(world).verify({ rendering: CONSISTENT, motion: PHYSICALLY_VALID, behavior: INTENT_ALIGNED });
Master's student in Computer Science at Brown University, working at the intersection of graphics, simulation, and embodied AI.
01 / Graphics Related Research & Development
A focused selection across reasoning, neural rendering, physical simulation, and real-time graphics.
Research · Brown University · Current
A portable, asset-level representation for articulated objects that encodes local rules such as latches, buttons, pedals, timers, interlocks, and coupled joints.
The system mediates simulator-proposed joint updates and supports LLM-assisted authoring, controller synthesis, and semantic-trace repair for evaluating embodied interaction tasks.
Physics Simulation · April 2026
A physics-based material point method sand simulator with Drucker–Prager elastoplasticity, return mapping, hardening, and friction-angle control.
Realtime Rendering · June 2026
An experimental Gaussian Splatting pipeline adapting ReSTIR-style reservoir resampling for visibility-aware reuse across pixels and frames.
Realtime Graphics · December 2025
A procedural environment renderer with fBm terrain, layered PBR materials, L-System vegetation, weather, water, fog, and instanced rendering.
02 / Research++
Work across clinical machine learning and generative imaging, with an emphasis on systems that can be tested rather than merely observed.
Medical AI · University of Rochester / URMC · 2024–2026
A Parkinson's-disease-specific machine learning model for six-month end-of-life prognostication among nursing home residents, evaluated against Medicare hospice guidelines.
AI-designed cover
Generative Imaging · Tsinghua AIR Lab · 2024
Investigating how Deep Contextual Video Compression and Diffusion Posterior Sampling can combine compression, denoising, and generative reconstruction.
03 / Experience
Software Development Engineer Intern
Jun–Aug 2023
Teaching Assistant & CSUG Tutor
Undergraduate
04 / More Work
Reinforcement Learning · 2024
A belief-aware agent for partially observable, stochastic grid combat.
Vision & Language · 2024
A reproducible VQA benchmark with ResNet/SBERT and CLIP/LoRA baselines.
AI-refined cover
History of Mathematics · 2024
A study of rectangular-array methods in The Nine Chapters on the Mathematical Art and their relationship to modern linear algebra.
Parameter-Efficient Fine-Tuning · Survey
A comparative study of LoRA variants, including AdaLoRA and LoRA-FA, evaluated on GSM8K and GSM-Plus.
05 / Tools
Small, focused utilities for everyday development and visual computing.
Classical Computer Vision · 2024
Sauvola binarization, geometric pose estimation, and Kalman-filtered tracking.
Ithaca, New York
06 / About
Hi, I'm Hanzhang — you can call me Henry. I am pursuing an Sc.M. in Computer Science at Brown University, advised by Daniel Ritchie. I earned dual B.S. degrees in Computer Science and Applied Mathematics from the University of Rochester.
My work tends to begin with the same question: how can we make intelligent systems easier to understand, test, and trust when they interact with complex visual or physical environments?