Hi, I'm Gaorui Zhang

I'm a PhD student at Zhejiang University working on neural decoding, scientific tooling, and reliable AI systems for research.

What I work on

• Neural decoding: EEG / sEEG, speech-related brain signals, cross-subject modeling, and brain signal understanding
• Scientific tooling: open-source tools for publication-ready figures, tables, and research workflows
• AI for research: semi-automated systems for literature review, project memory, and scientific assistance

Selected Research and Tools

These projects reflect my two main directions: neural decoding research and tools for reproducible, publication-oriented scientific workflows.

Research

speech-decoding · Paper · Repo · !Stars !Forks

Unified brain-to-text decoding across both speech production and perception, with one reproducible workflow spanning preprocessing, channel analysis, decoder training, LLM-based reconstruction, and post-hoc analysis.

NeuroSketch · Paper · Repo · !Stars !Forks

A neural decoding framework built through systematic architectural optimization, backed by 5,000+ experiments across modalities, signal types, and decoding tasks rather than isolated architecture tweaks.

Tooling

claude-scholar · Repo · !Stars !Forks

A semi-automated research assistant for literature review, coding, experiments, writing, and project memory, designed to keep human decision-making at the center instead of pretending to be a fully autonomous scientist.

pubfig · PyPI · Repo · !Stars !Forks

A Python library for publication-style scientific figures, focused on the last mile of paper figures: paper-ready defaults, clean export, and panel-first Figma assembly for multi-panel composition.

pubtab · PyPI · Repo · !Stars !Forks

A tool for bidirectional Excel↔LaTeX table conversion with stable roundtrip behavior, style preservation, and preview support, built for real paper workflows instead of one-shot export.

Current interests

• Cross-subject neural decoding
• Brain-to-text decoding
• Reliable agent systems for scientific research
• Reproducible and maintainable research software

Links

• Homepage: galaxy-dawn.github.io
• GitHub: @Galaxy-Dawn
• Email: [email protected]

Collapse