AI/ML engineers, software developers, and technical professionals interested in building intelligent, autonomous AI agents

1. Basic familiarity with Python
2. Understanding of API concepts
3. Basic knowledge of AI/LLMs is helpful but not required

Course Outline:

1. Introduction to AI Agents and Claude

Topics Covered

1. What are AI agents?
2. Agentic workflows vs traditional chatbots
3. Overview of Claude models
4. What is the Claude Agent SDK?
5. Typical use cases for Claude agents
   1. Assistants
   2. Workflow automation
   3. Research agents
   4. Internal copilots

Demonstration

Walkthrough of the SDK environment

2. Claude Agent Architecture

Topics Covered

1. Components of an AI agent
2. Claude model interaction patterns
3. System prompts and instructions
4. Tools and tool calling
5. Agent loops and reasoning cycles
6. Managing conversation state

Key Concepts

1. ReAct-style reasoning
2. Tool orchestration
3. Structured outputs
4. Context windows and memory

3. Creating Your First Claude Agent

Topics Covered

1. Installing the Claude Agent SDK
2. Project setup
3. Creating an agent instance
4. Writing system instructions
5. Running a basic agent loop
6. Handling user input and responses

Hands-On Lab

Build a simple Claude Agent that:

1. Answers questions
2. Performs reasoning steps
3. Produces structured responses

4. Designing Agent Behaviors

Topics Covered

1. Designing agent instructions
2. Handling multi-step reasoning
3. Controlling agent responses
4. Guardrails and constraints
5. Error handling

Hands-On Lab

Build an agent that:

1. Asks follow-up questions
2. Plans multi-step solutions

5. Integrating Tools and APIs

Topics Covered

1. What are tools?
2. Tool schemas and definitions
3. Calling external APIs
4. Using tools for:
   1. Search
   2. Databases
   3. Internal systems
5. Handling tool responses

Hands-On Lab

Build an agent that:

1. Calls a search API
2. Retrieves information
3. Synthesizes results

6. Knowledge and Retrieval for Agents

Topics Covered

1. Grounding agents with data
2. Retrieval-Augmented Generation (RAG)
3. Connecting agents to documents
4. Vector databases and embeddings
5. Improving answer accuracy

Hands-On Lab

Build an agent that:

1. Answers questions from a document set
2. Retrieves relevant context
3. Produces grounded responses

7. Testing, Observability, and Deployment

Topics Covered

1. Testing prompts and tools
2. Logging agent traces
3. Debugging agent behavior
4. Performance considerations
5. Deploying agents

Demonstration

Deploy an agent as a simple API service.

Training material provided: Yes (Digital format)