<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>Agentic Development on Aby George</title><link>https://abygeorgea.com/tags/agentic-development/</link><description>Recent content in Agentic Development on Aby George</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Sat, 19 Sep 2026 08:25:31 +1000</lastBuildDate><atom:link href="https://abygeorgea.com/tags/agentic-development/index.xml" rel="self" type="application/rss+xml"/><item><title>Shipping the Agent: Eval Costs, Free Tier Deployment, and the Chat UI</title><link>https://abygeorgea.com/blog/2026/09/19/shipping-a-resume-ai-agent-on-a-free-tier/</link><pubDate>Sat, 19 Sep 2026 08:25:31 +1000</pubDate><guid>https://abygeorgea.com/blog/2026/09/19/shipping-a-resume-ai-agent-on-a-free-tier/</guid><description>Measuring and cutting AI eval costs, deploying a RAG agent to a free tier without running out of memory, and embedding the chat UI in a Hugo blog.</description></item><item><title>How Do You Test an AI Agent? Unit Tests, Evals, and the Bugs They Found</title><link>https://abygeorgea.com/blog/2026/09/18/testing-an-ai-agent-with-pytest-and-promptfoo/</link><pubDate>Fri, 18 Sep 2026 08:25:31 +1000</pubDate><guid>https://abygeorgea.com/blog/2026/09/18/testing-an-ai-agent-with-pytest-and-promptfoo/</guid><description>A practical strategy for testing an LLM based agent with pytest and promptfoo, and the five real bugs the tests caught that manual checking missed.</description></item><item><title>Keeping a RAG Agent Honest: Grounding, Guardrails, and Prompt Injection</title><link>https://abygeorgea.com/blog/2026/09/17/grounding-and-guardrails-for-a-rag-agent/</link><pubDate>Thu, 17 Sep 2026 08:25:31 +1000</pubDate><guid>https://abygeorgea.com/blog/2026/09/17/grounding-and-guardrails-for-a-rag-agent/</guid><description>Structured output, a deterministic grounding check, a layered prompt injection defence, and API limits for a RAG agent, with the real bugs I hit.</description></item><item><title>Building an AI Agent That Answers Questions About My Resume</title><link>https://abygeorgea.com/blog/2026/09/16/building-a-resume-ai-agent-with-rag/</link><pubDate>Wed, 16 Sep 2026 08:25:31 +1000</pubDate><guid>https://abygeorgea.com/blog/2026/09/16/building-a-resume-ai-agent-with-rag/</guid><description>How I built a RAG agent that answers recruiter questions from my resume: the architecture, chunking, embeddings, Pinecone ingestion and retrieval.</description></item></channel></rss>