I work remotely from New Zealand, keeping overhead low and passing that on to clients.
I use open-source tools for my own work — the same tools I help clients implement. This means I understand them from daily use, not from reading documentation.
One engineer delivering what used to require a team. I've built a 6-agent AI development workflow — Planner, Builder, Reviewer, Tester, Deployer, and Docs — orchestrated by a custom Python daemon.
This isn't theoretical. It's how I built the platform you're looking at right now. The result:
Faster delivery with built-in quality gates at every stage
Consistent code review and testing through automated agents
26+ MCP server integrations for AI-tool interoperability
Multi-model consensus (Claude, GPT, Gemini) for architecture decisions