CHAPTERS:
00:00 - 02:58 Introduction
03:00 - 05:57 What is agent? Why does it need memory?
06:00 - 11:10 Memory implementation
11:18 - 14:28 Code Examples
14:29 - 28:30 DEMO (QueryWeaver)
29:00 - 33:42 Why Graph? (Scale, privacy)
34:00 - 55:10 Q&A
---
USEFUL LINKS:
1. https://github.com/FalkorDB/QueryWeaver
2. https://app.queryweaver.ai/
3. https://github.com/FalkorDB/falkordb
--
ABOUT THIS WORKSHOP:
People are upset their AI doesn't have enough personality. But what about stopping the agent from hallucinating and suffering from amnesia, where all the relationships and context between data points aren’t readily available to it?
Without that context, your agent can never make good recommendations. And without temporal understanding, yesterday's relevant information might be stale today - but your agent doesn't know that. So it's giving recommendations based on outdated context. No knowledge graph means no way to maintain connections or understand when information expires.
This is why we're doing this workshop about implementing memory that actually works. Graph-based memory with Graphiti and FalkorDB. Real solutions to stop your agents from making terrible recommendations and forgetting relationships the moment you hit that context limit!
ABOUT FalkorDB:
FalkorDB is an ultra-fast, multi-tenant graph database powering Generative AI applications and AI agent long-context memory.
