Hands-On AI: Implementing Agentic Systems

Course details

Agentic AI shifts the focus of programming from responding to user input to actively pursuing defined goals. While the programming languages and frameworks may be familiar, the structure of the programs and the infrastructure they use are very different. In this course, software developer and project manager Keith Casey explains and demonstrates these changes, exploring the architectural shifts before moving into three concrete projects, each with challenges for learners. The course concludes with a look at the future and the difficult problems—like privacy—that developers will need to figure out.

Learning objectives
Observe how program flow changes when agents are active.
Explore different approaches for integrating AI to the system.
Work through use cases, including GitHub issues, appointment scheduling, and travel booking.
Plan for privacy and transparency issues.

Concepts

Introduction

From idea to IDE - Your first agent starts here
What you should know

Intro to Agents

What are agents and agentic systems
Key aspects of great agents
Agents - Reactive vs. proactive
Agents in the real world

Overview of Agentic Systems

Agentic architectures and frameworks
Authentication and authorization for data
Authentication and authorization for actions
Decision-making 101 - Goals and prioritization
Decision-making 201 - Reasoning and uncertainty
Decision auditing
Use cases to avoid
Model Context Protocol (MCP)

Use Case 1 - Checking Issues on GitHub

Sketching out your use case
Connecting to your data sources
Connecting the pieces
Challenge - Filtering the noise on GitHub
Solution - Filtering the noise on GitHub

Use Case 2 - Scheduling an Appointment

Sketching out your use case
Connecting to your data sources
Connecting to your actions
Connecting the pieces
Challenge - Using your Slack workspace
Solution - Using your Slack workspace

Use Case 3 - Booking a Trip

Sketching out your use case
Connecting to your data sources
Connecting to your actions
Connecting the pieces
Improving your agent

Next Steps

Privacy and security in agentic systems
Auditing and transparency
Trying other data sources and actions
The future of agentic systems