Blog
Thoughts on DevOps, database engineering, and the art of automation.
What Is RAG? How Retrieval-Augmented Generation Works Internally
RAG is more than a vector database with an LLM. It's a structured architecture where an LLM retrieves external knowledge at runtime to generate grounded, accurate answers.
New Attack Surfaces in AI Systems: Understanding the Security Risks Unique to LLM Applications
AI systems introduce attack vectors that traditional security tools don't catch. Prompt injection, indirect injection, and tool misuse are now first-class security concerns.
Implementing Role and Group-Based Security with Qdrant in Enterprise AI Systems
Qdrant is a vector search engine, not an enterprise IAM system. Learn the correct architecture for implementing role and group-based access control in production RAG pipelines.
Why Security Matters in AI-Based Systems: Building Secure, Monitorable Enterprise AI Platforms
Many teams prioritize model quality over security architecture — a critical mistake. An AI system can unintentionally expose massive amounts of sensitive data in a single response.
How Data Is Stored and Retrieved from a Vector Database
Vector databases don't search for exact embedding values — they perform nearest-neighbor similarity searches. Here's the complete ingestion and retrieval flow explained.