Context & memory
What should an agent remember, and what should remain inspectable?
Jamie WeishaarI build software, lead engineering programs,
and design how people and AI work together.
Staff Engineering Program Manager at Netflix Founded Enterprise SRE
Explore selected workSelected work
Swift · macOS · MCP
Context you can inspect. Handoffs you can trust.


One local data model. A native app,
a CLI, and an MCP server.
Questions I’m exploring through building, reading, and practice.
What should an agent remember, and what should remain inspectable?
Where should automation act, and where should people stay in control?
How do we turn unexpected behavior into better systems?