Mainland China · Remote
中文English

About

Chengduo Xuan

AI / BCI / Engineering / Neuroscience

Building a long-term path across AI, brain-computer interfaces, and engineering. I want to connect learning, research, and real-world systems toward neuroengineering that can help people.

Introduction

I'm an independent learner moving toward neuroengineering. I treat learning, research, and engineering as one continuous practice rather than separate phases.

This site is both a public notebook and a quiet commitment device — what I publish here is what I am actually working on.

Research Direction

Brain–computer interfaces, with a focus on robust decoders, calibration-free models, and signals that survive real-world noise. I am especially drawn to systems that can eventually restore function for patients.

Learning Path

  1. Deepening BCI foundations

    2025 —
    • Signal processing, computational neuroscience, decoder architectures
    • Reproducing EEG decoding baselines end-to-end
    • Weekly reading log across neuroscience and ML
  2. Engineering practice

    2024
    • Local-first notes system, data pipelines, small ML systems
    • Rust and C for the parts where precision matters
    • Linux, shell, and reproducible workflows
  3. Starting the long path

    2023
    • First serious neuroscience reading list
    • Parallel deepening of math and programming
    • First Kaggle notebooks — habit over hype

Projects

Small projects I can finish, ship, and learn from: an EEG decoding pipeline, a literature notes app, a minimal spike sorter, and an ongoing daily learning log.

Technical Skills

Programming Languages

PythonTypeScriptRustCBash

Tools

LinuxNeovimtmuxGituvDocker

AI / ML

PyTorchNumPyscikit-learnTransformersNotebooks

BCI / Neuroscience

MNEEEGSpike sortingHodgkin–HuxleySignal processing

Engineering / Workflow

Local-firstReproducibilityDesign systemsTanStackReact

Long-term Goal

To do honest work at the intersection of neuroscience and engineering — work that helps real people, and that I can defend at every layer of the stack.