Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Juniper: Progressive Disclosure and Caching
Learn about custom orchestration techniques: progressive disclosure, string passing, and request level caching for improved agent performance and token efficiency.
I will be presenting the custom orchestration behind my app, Juniper. I will talk about the different techniques used related to context engineering and performance: progressive disclosure - a term coined by Anthropic engineers describing the process of exposing progressively lower level data to an agent as it narrows down what it needs, string passing - how my agents can pass string objects instead of regurgitating information, request level caching - how large pieces of data can be stored in a request level cache shared by all agents again to avoid regurgitation of large context.
Multi-agent AI assistant using Python/FastAPI, hybrid retrieval, and dynamic model switching.
Cross-platform AI assistant integrates wearable health metrics with productivity services.