The model is the easy part. What actually breaks AI projects is the layer underneath it. I focus on wiring agents into enterprise data, identity, security, and governance so they survive production.https://itssujeeth.com/https://itssujeeth.com/writing/01-claude-code-enterprise-setup/https://itssujeeth.com/writing/01-claude-code-enterprise-setup/How I configure Claude Code to work with real enterprise codebases. Covers CLAUDE.md patterns, context management, and the workflow habits that hold up when the codebase has 200k lines and a dozen services.Mon, 15 Jun 2026 00:00:00 GMThttps://itssujeeth.com/writing/02-mcp-server-patterns/https://itssujeeth.com/writing/02-mcp-server-patterns/A contrarian take on agentic system design and why thinking in Model Context Protocol tools first produces cleaner architectures than starting with an orchestration framework.Mon, 01 Jun 2026 00:00:00 GMThttps://itssujeeth.com/writing/03-rag-production-patterns/https://itssujeeth.com/writing/03-rag-production-patterns/After building enterprise RAG proofs of concept that stayed POCs, and then taking AI-assisted automation to production, here's what the architecture has to look like when the demo phase is over, and the mistakes I'd avoid the first time.Fri, 15 May 2026 00:00:00 GMT