AI architecture field notes
AI The Architect is a practical publication by Edy Figueroa on production AI systems, agent workflows, retrieval architecture, integrations, and governance.
Featured story
Large language models learn by comparing predictions against expected answers, adjusting internal parameters, and repeating that process across many examples.
Explore by topic
From the Lab
Short notes from building AI workflows, debugging retrieval, testing agent behavior, and turning rough demos into production systems.
Projects
Reusable patterns for AI platforms, WordPress and Django boundaries, data integrations, evaluation, and operational governance.
About Author
Edy Figueroa writes from hands-on work with AI platforms, agent workflows, RAG systems, MCP integrations, WordPress publishing surfaces, and Django-backed automation.
Latest articles
Large language models learn by comparing predictions against expected answers, adjusting internal parameters, and repeating that process across many examples.
What MCP servers do, how they expose tools, and how teams should govern them from local development through production rollout.
A governance checklist for proving that an agentic workflow is ready to ship, monitor, and improve after release.
A boundary map for using WordPress as a publishing surface, Django as the governed application runtime, and MCP as the tool integration layer.
How to design AI tool contracts that are small, permission-aware, testable, and useful for real enterprise workflows.
A practical explanation of MCP as the contract layer between AI agents, enterprise tools, permissions, and observable execution.