I build AI systems. Research first, then the system that meets the need. My own runs on persistent memory and rewrites its own playbook.
Agentic workflows, retrieval pipelines, and automation that runs in production rather than in a demo. Two years building a system that orchestrates my own work: persistent memory across sessions, automated capture and routing, processes that update themselves as I learn.
Deciding what to build and why, from problem framing through roadmap, scope, and the tradeoffs nobody wants to make.
Mixed-methods research and the data work behind it. Evidence that changes decisions, not decks that get filed.
The hard part isn't adding AI.
It's knowing which parts should never touch it.
Everything I capture becomes structure.
Meetings, documents, research, and voice notes route through one pipeline into a Postgres brain I host myself. It extracts the entities, compiles the wiki, and hands back the right file in zero tokens.
I came up through market research and data analytics, then digital marketing, where research met brand and web design. Today I work in information management and communications for a national research center. A master's in Information Management from UW pointed me at AI, and I've been building AI systems and teaching colleagues to do the same ever since.





If you're building with AI in health, science, research, or marketing, I'd like to hear what you're working on.
I'll reply from contact@amyyoung.me.