Building Scalable AI Memory for Agents Across Graphs and Vectors
Qdrant - Vector Database & Search Engine via YouTube
Learn AI, Data Science & Business — Earn Certificates That Get You Hired
Learn Backend Development Part-Time, Online
Overview
Google, IBM & Meta Certificates – 40% Off
One plan covers every Professional Certificate on Coursera.
Unlock All Certificates
Explore how to build scalable AI memory systems for agents using the cognee Python SDK in this 15-minute conference talk from Qdrant's Vector Space Day 2025. Learn to overcome common challenges when connecting company data to LLMs at scale, including data consistency issues, brittle contracts, microservice sprawl, and developer experience gaps. Discover how cognee abstracts storage and retrieval across multiple graph and vector databases while integrating seamlessly with various LLMs. Examine practical approaches to modeling, segmenting, and synchronizing memory across structured and unstructured data sources through a comprehensive finance example that combines documents, records, and relationships. Understand the design decisions behind multi-backend support and how embeddings and graph context work together to enhance AI agent capabilities. Master patterns for achieving low-latency retrieval with robust versioning, and explore strategies for modularity, testing, and developer experience that reduce system fragility while accelerating iteration cycles. Gain insights into what "AI memory" truly means for agents, how to provide durable and relevant context, and how to evolve your architecture without vendor lock-in.
Syllabus
Building Scalable AI Memory for Agents Across Graphs and Vectors | Cognee | Vasilije Markovic
Taught by
Qdrant - Vector Database & Search Engine