Location: Gig Harbor, WA, USA
Remote: Yes (preferred)
Willing to relocate: Yes (Bay Area), possibly elsewhere for the right role
Technologies: TypeScript, Node.js, React, Three.js / Babylon.js, MySQL, MCP, multi-LLM orchestration, Capacitor
Résumé/CV: https://1drv.ms/b/c/72423309fdf980dc/IQBeEr4_o85zQrE1OzyCtTo...
Email: robertgardunia@hotmail.com
Staff systems architect, 30 years shipping production. I architect with AI agents as the build substrate, not autocomplete — running parallel agents per subsystem, coordinating through a shared memory layer I built.
Currently focused on AI infrastructure: memory2thought (patent-pending cross-LLM persistent memory across Claude/ChatGPT/Gemini/Grok, live at memory2thought.com) and thought2speech (patent-pending atypical speech recognition for Down syndrome speakers).
Recent work includes a municipal civic engagement platform shipped solo in 12 weeks (cgprojecttracker.com), a self-healing distributed fleet management system (DFF) for educational gaming installations, and Visual DNA — a privacy-preserving identity system for K-12 learners.
Looking for senior architect or AI infrastructure work where I can own end-to-end delivery: applied AI products, multi-agent systems, MCP-based architectures, or anything where deep systems thinking compounds across the stack. Open to fractional CTO, advisor, or principal engineer roles.
Staff systems architect, 30 years shipping production platforms. Currently focused on AI infrastructure.
Recent work:
> memory2thought (memory2thought.com) — Patent-pending cross-LLM persistent memory across Claude, ChatGPT, Gemini, and Grok. Live with paying users.
> thought2speech — Patent-pending atypical speech recognition for Down syndrome speakers. 2.7x accuracy improvement over baseline at patent filing.
> City Project Tracker (cgprojecttracker.com) — Municipal civic engagement platform shipped solo in 12 weeks. Six integrated subsystems including a Three.js AI avatar with ARKit blend shapes, geospatial project tracking, and sentiment analytics.
> DFF — Self-healing distributed fleet management for educational gaming installations. Concept to working 29-node federation in four days.
> vDNA — Privacy-preserving identity system for K-12 learners.
Looking for senior architect or AI infrastructure work where I can own end-to-end delivery: applied AI products, multi-agent systems, MCP-based architectures, or anything where deep systems thinking compounds across the stack. Open to fractional CTO, advisor, or principal engineer roles.
Staff systems architect, 30 years shipping production. I architect with AI agents as the build substrate, not autocomplete — running parallel agents per subsystem, coordinating through a shared memory layer I built.
Currently focused on AI infrastructure: memory2thought (patent-pending cross-LLM persistent memory across Claude/ChatGPT/Gemini/Grok, live at memory2thought.com) and thought2speech (patent-pending atypical speech recognition for Down syndrome speakers).
Recent work includes a municipal civic engagement platform shipped solo in 12 weeks (cgprojecttracker.com), a self-healing distributed fleet management system (DFF) for educational gaming installations, and Visual DNA — a privacy-preserving identity system for K-12 learners.
Looking for senior architect or AI infrastructure work where I can own end-to-end delivery: applied AI products, multi-agent systems, MCP-based architectures, or anything where deep systems thinking compounds across the stack. Open to fractional CTO, advisor, or principal engineer roles.
Portfolio: knightsrook.com / noshit.software