Independent engineer / open source developer
I build practical systems where automation, artificial intelligence, and scientific computing meet. My work favors clear interfaces, traceable data, and results that another engineer can inspect and rebuild.
GitHub · Interactive edition · Source · Deployment notes · 2026
Research infrastructure · Repository
The working home for multi-source research data acquisition and synchronization in climate and Earth-science workflows. Its problem space spans source routing, download-chain verification, resumable tasks, and data integrity checks.
Scientific computing · Repository
A Tibetan Plateau cloud-water experiment system built around proxy indicators, a GCWR water-budget model, three-layer error analysis, reproducible runs, and scientific figures. HydroWeave-DBAS is a separate data-management system rather than another analysis layer inside this repository.
Data systems · Repository
A local data-asset system for ERA5 and NetCDF volumes: cataloging, readability checks, SHA-256 verification, online backup, resumable task orchestration, and a query-aware Zarr layer. The repository explicitly keeps experiment analysis and plotting out of scope.
Modeling toolchain · Repository
MCMtools is a local-first Python toolchain for mathematical-modeling competitions. Its working scope covers structured project state, experiment runs, reusable figure and table generation, paper builds, and a consistency audit that runs before compilation. The roadmap remains separate from what is implemented.
The descriptions below are working directions, not proficiency ratings. They distinguish current practice from interests that I am still developing.
01 / AI Engineering
Status: developing practice. Agent systems, orchestration, and tool-using workflows.
02 / Automation
Status: project evidence. Multi-source acquisition, resumable tasks, integrity checks, and task orchestration.
03 / Scientific Computing
Status: project evidence. Cloud-water modeling, ERA5/NetCDF data systems, verification, and reproducible analysis.
04 / Embedded Systems
Status: developing practice. Control, embedded systems, FPGA study, and edge intelligence; not a production-experience claim.
05 / Software Engineering
Status: project evidence. Modular boundaries, tests, CI, documentation, and auditable delivery in public repositories.
01 — Reproducibility
A result should be traceable, repeatable, and independently verifiable.
02 — System Thinking
Solve problems through well-defined components and reliable workflows.
03 — Evidence over Assumptions
Test the actual system. Measure performance. Analyze failures.
04 — Simplicity over Complexity
Build only what is necessary, but build it properly.
- Autonomous AI agents and multi-agent systems.
- Scientific data infrastructure.
- Edge intelligence and embedded AI.
- Mathematical modeling and optimization.
- Reproducible engineering experiments.
GitHub is the public contact route: @Neiyako.
For project-specific questions, open an issue in the relevant repository so the discussion stays with the code and evidence.
The interactive edition is a Three.js engineering notebook in this repository. The deployable source begins at index.html.
The live edition is a Three.js engineering notebook built with Vite, TypeScript, and the WebGPU renderer. The production build is published through GitHub Actions at neiyako.github.io/Neiyako. Deployment details and repeatable verification steps remain in DEPLOYMENT.md.
| Surface | Status |
|---|---|
| Interactive edition | Live and verified at desktop and mobile widths |
| Deployment runbook | Available in DEPLOYMENT.md |
Build. Experiment. Verify. Improve.
NEIYAKO · ENGINEERING NOTEBOOK · 2026

