<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Chris Nielson · writing</title><description>Chris Nielson is a software engineer in Brooklyn, NY building AI products: production agents, knowledge systems, and local inference.</description><link>https://cjnl.dev</link><item><title>GraphRAG as Scoped Context for Production Agents</title><link>https://cjnl.dev/writing/knowledge-graphs-vs-rag</link><guid isPermaLink="true">https://cjnl.dev/writing/knowledge-graphs-vs-rag</guid><description>How knowledge graphs can reconstruct small, permission-aware context packets for agents, reduce context rot, and preserve relationships, provenance, and time.</description><pubDate>Sat, 13 Jun 2026 00:00:00 GMT</pubDate><category>GraphRAG</category><category>AI Agents</category><category>Knowledge Graphs</category><category>Context Engineering</category><category>Agent Memory</category></item><item><title>Semantic Models for Databases</title><link>https://cjnl.dev/writing/semantic-models-for-databases</link><guid isPermaLink="true">https://cjnl.dev/writing/semantic-models-for-databases</guid><description>Why semantic models are becoming the missing layer between raw database schemas, business meaning, analytics, and AI systems.</description><pubDate>Thu, 23 Apr 2026 00:00:00 GMT</pubDate><category>Data</category><category>Databases</category><category>Semantic Models</category><category>AI</category><category>Risk</category></item><item><title>Production AI Agents in Enterprise Systems</title><link>https://cjnl.dev/writing/production-ai-agents-in-enterprise-systems</link><guid isPermaLink="true">https://cjnl.dev/writing/production-ai-agents-in-enterprise-systems</guid><description>A practical architecture for building AI agents that can work with tools, data, permissions, and human review in production environments.</description><pubDate>Thu, 09 Apr 2026 00:00:00 GMT</pubDate><category>AI Agents</category><category>Software Engineering</category><category>Enterprise</category><category>Architecture</category><category>Risk</category></item></channel></rss>