Memory & State Management in LLM Applications

How do we build AI applications that remember information, maintain context, and pick up where they left off?
In this session, we'll explore memory and state management in LLM-powered applications, from maintaining conversation history to persisting information across sessions and managing complex workflows.
We'll look at the architectural patterns and engineering decisions behind building AI systems that maintain continuity beyond a single interaction.

What we'll cover:

Understanding the difference between context, memory, and state
Managing conversation history and context windows
Short-term vs. long-term memory
Memory storage, retrieval, and summarization strategies
State persistence, checkpointing, and resuming workflows
Building an LLM application with memory and persistent state

Through a hands-on example, we'll explore how these concepts work in practice and discuss the trade-offs involved in designing memory-enabled AI applications.

Who should attend?
Software engineers, AI enthusiasts, and anyone interested in building practical LLM-powered applications. Basic familiarity with LLMs and programming is recommended.