RetroGhostLabs

Solo Developer & Content Creator

Building indie games and sharing AI-accelerated development workflows. Creating YouTube content about developer productivity and releasing open-source templates for AI-assisted development.

View Projects YouTube GitHub

About Me

I'm Shaun, a solo developer & content creator at RetroGhostLabs. I build indie games and share AI-accelerated development workflows. I create YouTube content about developer productivity and release open-source templates for AI-assisted development.

What I Do

I build games, create content, and design AI-powered workflows that let solo developers move fast without burning out. My work focuses on:

  • Local LLM infrastructure
  • Dual-AI development setups (Cline + Claude CLI)
  • Knowledge management for rapid iteration

Templates & Resources

My GitHub includes open-source templates and resources:

  • Obsidian Vault Templates (PARA + Zettelkasten)
  • AI Prompt Libraries
  • Game Design Documents & Templates
  • Workflow Automation Scripts
  • IDE Configurations

On the Side

When I'm not coding or recording:

  • Writing music & sculpting tone chains
  • Tuning smart home setups
  • Exploring retro tech & creative storytelling

Current Projects

Game Development

Electorate

Turn-based political strategy game with procedural elements

  • Platform: PC (Steam target)
  • Engine: Unity
  • Timeline: 11-month cycle (Week 3)
  • Tech: C# + Unity, AI-generated assets
Side Project

Pooping Pugs: Unleashed

Comedic mini-game side project for creative breaks

  • Exploring 3D modeling with Meshy.ai
  • Roblox prototype → Unity rebuild
  • Experimental & playful approach
Content Creation

YouTube: RetroGhostLabs

Content for developers who want to build faster with AI

  • AI Workflow Engineering
  • Solo Dev Systems & Dev Logs
  • Tools & Setup Reviews
  • Template Releases

Stack: Obsidian, Final Cut Pro, OBS

Visit Channel →

Infrastructure

Specter LLM Server

Local GPU-accelerated LLM inference for cost-effective AI development

  • NVIDIA RTX 4080 SUPER, 128GB RAM
  • 100-130 tok/s on 7B-13B models
  • Saves $1,500-$3,300 annually vs. cloud
  • Stack: Ollama, 15+ optimized models
Developer Tools

Development Infrastructure

Sophisticated multi-machine environment with heavy automation

  • Dual-AI: Cline + Claude CLI
  • Auto-detection of framework rules
  • Memory Bank for shared context
  • One-command Makefile init

Coming soon: Public dotfiles repo

Tech Stack

Game Development

  • Unity (primary), Godot (exploring)
  • C#, JetBrains Rider
  • Midjourney, Meshy.ai (assets)

AI Development

  • Claude CLI, Cline, ChatGPT
  • Ollama (self-hosted)
  • Dual-AI workflow automation

Backend & Infrastructure

  • TypeScript, Node.js, Python
  • Docker, PostgreSQL, MongoDB
  • Git-driven development

Content & Knowledge

  • Obsidian (20+ plugins)
  • Final Cut Pro, OBS Studio
  • Markdown-first documentation

My Philosophy

Build Fast, Iterate Faster

Heavy AI assistance for acceleration. Prototype first, polish later. Ship working code over perfect code.

Self-Hosted > Cloud Dependencies

Local LLM server saves thousands annually. No rate limits, no API costs, full control over infrastructure.

Document Everything

Markdown-first docs that feed back into AI tools for better context. Everything version controlled.

Playful & Scrappy

Humor + tech + nostalgia. Slightly chaotic but organized underneath. Creative experimentation encouraged.

Get In Touch

Have questions? Want to collaborate? Interested in templates or workflows?

headghoul@retroghostlabs.com