Talk

AI apps are great, but AI apps that work everywhere, offline, and with (almost) no latency are even better.

In this talk, we'll explore how to build a simple chat app with conversational memory, a knowledge base, RAG (allowing AI to search your documents for information), tools (enabling an agent to execute code), and an Agent (letting the LLM make decisions), that works on both mobile and desktop.

This will be an opportunity to introduce you to the following tools:

• Couchbase Lite: a lightweight database that supports SQL and vector search.
• Google LiteRT: models optimized for mobile.
• Tauri: a framework for creating desktop and/or mobile apps.

You'll leave this talk with the essential knowledge to build your first edge and AI apps, and an understanding of the current limitations of small AI models.