Trent Doney

I’m interested in production-grade AI agent infrastructure: operational memory, multi-agent orchestration, MCP/Skills tooling, retrieval with provenance, and local-first AI systems. I build around PostgreSQL/pgvector, TypeScript, Python, and self-hosted inference, with a practical focus on agents that can remember, recover, and produce inspectable evidence beyond demos. I’m also interested in creative AI pipelines, including ComfyUI, LoRA workflows, image/video generation, and using generated media as proof of real system capability.