Skip to content
View f0909172434's full-sized avatar

Highlights

  • Pro

Block or report f0909172434

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
f0909172434/README.md

Chih-Kai Wang | 王治凱

I build Python and TypeScript tools for reliable testing, interactive applications, and AI-assisted research.

I am a B.S. student in the Mathematics Education Division at National Taipei University of Education, expected to graduate in 2028. I am based in Taipei and looking for software engineering and AI application internships.

Portfolio · CV · Email

Selected projects

Project What I built Explore
Finite Witness A browser workbench that searches finite graphs for counterexamples. People and agents use the same search engine through the UI and eight WebMCP tools. Try it · Source
HonestCI A TypeScript CLI and GitHub Action that checks whether JUnit test reports are present, fresh, non-empty, and consistent with a trusted baseline. Demo · Source
RigorGraph A Python CLI that connects research claims to evidence, checks record integrity, and generates a self-contained offline report. Open report · Source
ProofWeave Core Experimental Python tools for checking structured mathematical claims with Lean and retaining inspectable certification artifacts. Actual run · Source
SAIR Proof Press A public companion to Lean-checked solvers for equational implication, with frozen artifacts, evaluation summaries, and an English research paper. Project site · Source
MiniHarness 38 Traditional Chinese lessons, an eight-step harness workshop, quizzes, assignment checks, and a trainable tiny Transformer. Interactive demo · Workshop

Engineering case studies

Finite Witness · HonestCI · RigorGraph · ProofWeave · SAIR Proof Press · MiniHarness · External Windows contribution

Open-source contribution

I contributed a merged Windows verification fix to Codex Dream Skin. The change narrowed native-window fallback handling so unrelated errors remained failures, and repaired helper loading in standalone verification.

Tools and interests

  • Python, TypeScript, JavaScript, GitHub Actions, JSON Schema, and MLX.
  • Lean 4 / Mathlib: developing through proof-checking and research projects.
  • AI for mathematics, finite counterexample search, developer tools, and reproducible evaluation.

I also work on statistical procedure-selection experiments and source-retaining search tools.

My project reports keep measured results and limitations together. A finite graph search has a declared bound, a Lean certificate applies to its exact formal target, and experimental model results retain their negative findings.

For internships or collaboration, contact f0909172434@gmail.com.

Pinned Loading

  1. finite-witness-webmcp finite-witness-webmcp Public

    Interactive finite-graph counterexample search with inspectable certificates and eight shared WebMCP tools.

    JavaScript

  2. honest-ci honest-ci Public

    Make green CI mean the tests you expected actually ran.

    TypeScript 3

  3. rigorgraph rigorgraph Public

    Local-first claim-evidence graphs and deterministic audit reports for AI-assisted research.

    Python 1

  4. proofweave-math-lab proofweave-math-lab Public

    Experimental Python and Lean tools for structured claim certification. Core 2.0.0; evidence release v0.1.0.

    Python 2

  5. sair-stage2-proof-press sair-stage2-proof-press Public

    Public companion to Lean-checked equational implication solvers: frozen artifacts, released-input evaluations, and an English research paper.

    TypeScript 1

  6. miniharness miniharness Public

    Python agent-harness tutorial, eight-step workshop and interactive loop demo. Broader curriculum in progress: 2 ready / 38 modules.

    Python