<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Building AI Agents on DATATWEETS</title><link>https://datatweets.com/courses/generative-ai/building-ai-agents/</link><description>Recent content in Building AI Agents on DATATWEETS</description><generator>Hugo</generator><language>en</language><copyright>Copyright (c) 2025 Datatweets</copyright><lastBuildDate>Sat, 27 Jun 2026 09:00:00 +0200</lastBuildDate><atom:link href="https://datatweets.com/courses/generative-ai/building-ai-agents/index.xml" rel="self" type="application/rss+xml"/><item><title>Lesson 1 - What Makes an Agent</title><link>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-1-what-makes-an-agent/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-1-what-makes-an-agent/</guid><description>An agent is a model in a loop with tools, deciding its own steps until a goal is met. Learn the think-act-observe cycle, how agents differ from scripted workflows, and see a real agent chain two tools on its own.</description></item><item><title>Lesson 2 - The Agent Loop</title><link>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-2-the-agent-loop/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-2-the-agent-loop/</guid><description>Turn a single model call into an agent. Write a reusable agent loop that checks the stop reason, executes requested tools, feeds results back, and repeats until end_turn — with a max-steps cap so it can never run away.</description></item><item><title>Lesson 3 - Giving Agents Memory and Tools</title><link>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-3-giving-agents-memory-and-tools/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-3-giving-agents-memory-and-tools/</guid><description>Make retrieval a tool the agent chooses to use, instead of always forcing it. Back a search_docs tool with a vector database, let the agent decide when to call it, and keep conversation memory so follow-up questions just work.</description></item><item><title>Lesson 4 - Planning and Multi-Step Tasks</title><link>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-4-planning-and-multi-step-tasks/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-4-planning-and-multi-step-tasks/</guid><description>Real tasks need more than one step. Use a system prompt to encourage the agent to plan and use tools, give it a multi-step goal, and watch it chain several tool calls — then learn the practical limits of agent planning.</description></item><item><title>Lesson 5 - Guided Project: Research Assistant</title><link>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-5-guided-project-research-assistant/</link><pubDate>Fri, 14 Nov 2025 09:00:00 +0200</pubDate><guid>https://datatweets.com/courses/generative-ai/building-ai-agents/lesson-5-guided-project-research-assistant/</guid><description>The capstone: combine the agent loop, multiple tools, a vector-database search tool, memory, and a planning system prompt into a research assistant that decides which tools to use, chains them across steps, and answers real multi-step questions from your docs.</description></item></channel></rss>