I build things that run locally: voice agents, desktop apps, and automation tools.
Go · TypeScript · Python · Wails · Whisper · React
I like the hard layer most people skip — fully local AI inference, native desktop apps, and tools that work with zero cloud dependency. Everything below runs on my machine, offline, right now.
🎙️ JARVIS — Local Voice Assistant
Iron-Man-style desktop voice assistant. 100% local, zero cloud, zero subscriptions.
whisper.cpp(tiny.en GGML) for on-device speech recognition- Pure-Go intent router with fuzzy matching — 160+ sites, 40+ apps, filesystem indexing
- Windows SAPI voice output, live HUD in React
- Built with Wails v2 (Go + WebView2)
- Speech never leaves the machine — audio is captured in Go, transcribed on localhost
🌊 wavecursor — Gesture Cursor Control
Control your computer cursor with hand gestures in front of your webcam.
- Computer-vision hand tracking
- Built on Wails v3 (Go) — cross-platform desktop app
- No extra hardware; just your webcam
🗑️ cache_crusher — Junk File Crusher
Desktop app that hunts down and crushes junk files to reclaim disk space.
- Fast scanning engine, safe deletion
- Built with Wails + TypeScript
🦈 sharknote — Obsidian-style Notes Desktop App
Notes app with a knowledge graph, bidirectional links and export. Wails v3 + Go + React + SQLite, styled with shadcn/ui.
🔍 web-tech-detector — Website Tech Stack Scanner
Detect the technologies behind any website and generate clean HTML reports.
| Area | Tools |
|---|---|
| Languages | Go, TypeScript, Python, JavaScript |
| Desktop | Wails (v2/v3), WebView2, native Windows APIs |
| AI / ML | whisper.cpp, GGML, local LLM inference, SAPI |
| Frontend | React, Vite |
| Systems | Windows internals (winmm, WM_CLOSE, process control), CI |
Currently: building AI agents that run on-device, hunting for a remote role where I ship Go/TypeScript daily.