Xueyang Song

Dark Data Observatory

Xueyang Song

I work at the intersection of electrochemical materials, atomic-layer deposition, and applied machine learning. I also build research tools that keep scientific workflows local, inspectable, and reproducible.

zinc interfacesALD coatingselectrolytessmall-data MLlocal-first tools

Material Atlas / Zn Interface System

Nodes connect research questions, evidence, and local-first tools.

Zn anode Research domain Electrolyte Research domain ALD coating Research domain ML models Software tool Literature corpus Evidence layer Desktop tools Software tool Dataset atlas Dataset/project

Zn anode

Interfacial stability, coating behavior, and lifespan signals.

Electrolyte

Water-in-salt formulations, solvents, windows, and conductivity.

ALD coating

Protective surface layers for battery interfaces.

Scholar citations

672

Scholar h-index

6

Public GitHub repos

16

Atlas regions

What I work on.

The map connects the same research surface from different angles: materials, evidence, models, and software.

domain 01

Aqueous zinc batteries

I study interfacial stability, dendrite suppression, water-in-salt electrolytes, and practical screening of zinc-ion battery materials.

Zn metal
Zn metalelectrolytesinterfaces
domain 02

ALD and materials coatings

I use atomic and molecular layer deposition to think about protective coatings, anode stabilization, and surface-controlled battery behavior.

ALD
ALDMLDsurface chemistry
domain 03

Scientific ML software

I build around small-data modeling, literature curation, notebook workflows, and local-first desktop tools for research teams.

ML
MLElectronlocal-first

Software layer

Local-first tools connected to the research map.

View all apps
Literature signals flow into project-scoped artifacts.
Local Traceable Approved

Paper Pilot

coming soon

I'm building Paper Pilot as a local-first Electron research assistant for crawling papers, organizing artifacts, and generating AI-backed scientific briefs.

Product page
Notebook context stays local before model guidance leaves the workspace.
Local Traceable Approved

AcademiaML

coming soon

I'm building AcademiaML as a notebook-first Electron desktop tool for researchers with tabular data who need help selecting and running ML workflows.

Product page

Evidence layer

Selected cited papers.

Publications appear as evidence layers in the atlas: citation trails, material systems, and interface questions.

Dataset and project layer

Research code I keep traceable.

Battery Paper Crawler

A targeted, auditable Scholar-based literature harvesting workflow for battery and electrochemistry research.

literaturePythoncrawler

Radiative Cooling Workspace

A machine-learning-assisted workflow that links candidate proposal, model comparison, and physics re-verification.

materials MLverificationReact