AI Agent Fundamentals on AWS

Description

This two-day, instructor-led course teaches developers how to build their first AI agents on AWS, from a single model call to a deployed, observable agent. Participants write the model-and-tool loop by hand with the Amazon Bedrock Converse API, then rebuild it with the open-source Strands Agents SDK, connect tools through the Model Context Protocol, and choose between a fixed workflow and an agent for each task. They add memory and a knowledge base, protect the agent with Amazon Bedrock Guardrails, least-privilege permissions, and human approval for risky actions, then deploy it to Amazon Bedrock AgentCore Runtime and read its traces and evaluation results. The capstone assembles these parts into one working agent for a business task. About three-fifths of the course is hands-on, with guidance that decreases from step-by-step early labs to an independently assembled capstone. Every lab runs in a hosted AWS sandbox and uses foundation models on Amazon Bedrock. The course is the first in a three-course ladder: LF-20302 then takes agents to production on AgentCore, and LF-20304 covers enterprise agent architecture.

Who This Course Is For

• Software developers who want to add AI agents to their applications
• Cloud and DevOps engineers who will build, deploy, or support agents on AWS
• Data and ML engineers moving from single model calls to tool-using agents
• Technical leads who need hands-on grounding before choosing an agent approach

What You Will Learn

• Explain how an AI agent works and choose between a single model call, a fixed workflow, and an agent for a task.
• Build tool-using agents with Strands Agents, connecting tools written in code and tools supplied through the Model Context Protocol.
• Give an agent memory and knowledge, and protect it with guardrails, least-privilege permissions, and human approval for risky actions.
• Deploy an agent to Amazon Bedrock AgentCore, observe and evaluate its behavior, and show that it meets its requirements.

Prerequisites

• Working Python: functions, packages, virtual environments, and reading JSON
• Basic AWS experience: signing in to the console, IAM roles and policies, and running AWS CLI commands
• Familiarity with what a large language model and a prompt are; no prior agent experience is needed
• A modern web browser and a reliable internet connection; the lab sandbox is provided

Delivery Options

• Live, instructor-led
• Virtual instructor-led