For the data engineering crowd, here’s an explainer of how unstructured AI memory works, though the lens of what we know from working with structured data.

dltHub

An explainer bridging structured data engineering concepts (canonical models, dimensional models, semantic layers) with unstructured AI memory systems. It maps familiar data warehouse patterns to knowledge graphs, showing how subjects/objects/relationships in a graph mirror fact/dimension tables. Cognee, a Python SDK, is used as a concrete example: its four core operations (add, cognify, memify, search) parallel classic ETL pipeline stages, combining graph, vector, and relational stores to build persistent, self-improving memory for AI agents.

AI Memory: Understanding Modeling for Unstructured Data

A primer into structured data modeling Link icon

How This Translates to Unstructured Data Link icon