Members-Only
Recent Talks & Demos are for members only
You must be an AI Tinkerers active member to view these talks and demos.
Context Engineering: Sessions and Memory
Learn context engineering to build stateful AI agents using Sessions for short-term history and Memory for long-term personalization across conversations.
As AI agents move beyond single-turn interactions, context engineering becomes critical to building experiences that feel coherent, personalized, and truly intelligent.
In this talk, we explore context engineering as the practice of dynamically assembling and managing information within an agent’s context window to enable stateful, multi-turn AI systems. We introduce two foundational concepts: Sessions, which capture the short-term history of an ongoing conversation, and Memory, which provides long-term persistence across conversations.
Through hands-on codelabs using the Agent Development Kit (ADK), you’ll learn how to:
Manage conversation history within a session to create coherent, multi-turn interactions
Design and control working memory so agents can reason with evolving context
Extend agents with long-term memory that persists across sessions, enabling personalization over time
By the end of the session, you’ll have a practical framework, and working code, for building AI agents that don’t just respond, but remember, adapt, and evolve.