Note
This profile was summarized by ChatGPT from my GitHub repositories. It may contain simplifications or imperfect interpretations of my work.
AI research · optimization · agent infrastructure · browser experiments
https://zyh.sryze.cc/
GitHub Pages fallback: https://hunluanzhizhu.github.io/
My personal site is not a conventional portfolio page. It is a browser-native laboratory where research artifacts, interactive experiments, graphics, games, AI-assisted creations, and small applications are published as things you can actually open and use.
The site itself is deliberately lightweight: a static, framework-free front page with inline CSS / JavaScript / fonts, while individual experiments live under /projects/.
|
Rust · Bevy · WebAssembly A browser-based 3D sandbox experiment compiled to WASM. |
Slides · KaTeX · SVG · WebGL Research presentations published directly as browser experiences, with a dedicated session index. |
|
Benchmark · Godot · WASM A browser-published evaluation of AI-generated games and game-building workflows. |
TensorFlow.js · SNN · Local AI A browser-local ECG / AI experiment exploring client-side inference and neural modeling ideas. |
|
AI · Video · Canvas An interactive calibration experiment combining AI-generated material, video, and browser graphics. |
WebGL2 · Shader · Interactive Design A browser graphics experiment centered on procedural visuals, interaction, and shader-driven effects. |
The homepage is also part of the experiment: project filtering and search, Chinese / English switching, procedural visuals, WebGL effects, motion controls, responsive layouts, keyboard navigation, and small interaction details are all implemented directly in the page.
|
Optimization algorithms, reinforcement learning, spiking neural networks, paper reproduction, experimental model ideas. |
Reusable agent skills, automated research / coding workflows, model API infrastructure, workspace tooling. |
WebAssembly, WebGL, Canvas, SVG, browser-native demos, games, interactive research presentations. |
paper → mechanism → reproduction → modification → experiment → engineering → automation
|
Adaptive fractional gradient-descent optimizers with PyTorch integration and a spiking-neural-network framework.
|
A lightweight Rust proxy for OpenAI-style, Anthropic-style, and Responses-style model APIs, with streaming, retries, mapping, and request adaptation.
|
|
My reusable AI-agent skills and workflow layer: behavior rules, skill organization, version locking, and research / productivity tooling.
|
An experiment around automating research and coding workflows with AI agents.
|
|
AI-assisted game creation and evaluation work, connected to the public browser demos in my Browser Lab.
|
An interactive AI-related calibration project also published as a live browser experience.
|
The website is the visual surface of that process; the repositories hold the underlying algorithms, agent workflows, infrastructure, and experiments.
Mirrors, clones, and non-representative repositories are intentionally excluded from the core-project summary.