Architecture

Enterprise architecture for AI systems: tools, identity, retrieval and reversibility, with the evidence behind each design choice.

Most agent pilots stall on the architecture around the model, not the model. What surveys, published incidents and reliability benchmarks show about tools, identity, retrieval, reversibility and measurement, and the six questions I ask before anything ships.

Ivan Bardak

Oct 29, 20266 min read

GraphRAG vs RAG: what the evidence actually says

GraphRAG wins on multi-hop and whole-corpus questions and loses on plain lookups, at a higher cost. What three studies show, how much weight each deserves, and a decision rule you can use.

Ivan Bardak

Oct 29, 20266 min read

Five questions before you build a knowledge graph for RAG

A knowledge graph can make RAG better at connecting facts, or just slower and more expensive. Five questions, each tied to published evidence, that tell you which one you are about to get.

Ivan Bardak

Oct 29, 20266 min read

MCP tool design: fewer, safer, testable tools

Accuracy falls as the tool list grows, and every major vendor now says keep it small. What the MCP specification, the vendors and independent benchmarks say about tool design, and the checklist I use to review an MCP server.

Ivan Bardak

Oct 29, 20266 min read

Delegated access for agents: OAuth 2.1, PKCE and scoped tokens

On whose behalf is the agent acting, and with what rights? The delegation model I use, what the MCP authorization specification requires as of 2026, how to design scopes, and where agent credentials actually leak.

Ivan Bardak

Agentic AI from demo to production: what the evidence says about the architecture

GraphRAG vs RAG: what the evidence actually says

Five questions before you build a knowledge graph for RAG

MCP tool design: fewer, safer, testable tools

Delegated access for agents: OAuth 2.1, PKCE and scoped tokens