AI Agents · TRACK 01/05
Agent Fundamentals
What an agent actually is, how the loop runs, and whether you need one.
// THE TRACK
01 · START HEREIntro to AI AgentsUnderstand what turns an LLM into an agent — the loop of reasoning, tool calls, and feedback that lets a model act instead of just answer.BEGINNERThe ReAct PatternUnderstand the Thought → Action → Observation loop that powers most LLM agents, why it outperforms plain chain-of-thought, and where it breaks down.INTERMEDIATEThe Agent LoopWalk through one full iteration of the loop at the heart of every agent and see exactly what the model receives and returns at each step.BEGINNERAgents vs Workflows vs ChatbotsLearn the dividing line between scripted workflows, conversational chatbots, and agents that choose their own next step.BEGINNERAgentic Workflow PatternsLearn the five named workflow patterns that cover most real use cases — and how to combine them before reaching for a fully autonomous agent.INTERMEDIATEAgent Decision FrameworkGet a simple decision framework for choosing between a single prompt, a fixed workflow, and a full agent — and avoid overbuilding.INTERMEDIATEEvaluating AgentsLearn why a single accuracy number can't measure an agent, and which task-level, step-level, and cost metrics actually tell you if it works.INTERMEDIATEAgent Failure ModesRecognize the recurring ways agents break — loops, derailment, hallucinated actions, and compounding errors — so you can design around them before they bite.INTERMEDIATEAgent Cost & LatencyUnderstand exactly what makes agents slow and pricey — repeated calls over a growing context — and the concrete levers that cut both.INTERMEDIATE