Build with AI: Create an Agent with GPT-OSS

Course details

Get up and running with building your own AI planner agent, working with code in the first 15 minutes. In this project-based course, join Laurence Moroney—award-winning AI researcher, best-selling author, and fellow at the AI Fund—as he outlines a build-along approach and guides you through the steps of creating an outdoor activity planner. Get started with a simple "Hello World" agent, before progressively adding weather parsing, web scraping, and intelligent decision-making capabilities using open-source LLMs. An ideal fit for developers looking to prioritize data privacy, costs, and customization, this course focuses on practical implementation over theory, highlighting agentic design patterns through real code that you can immediately test and iterate on. Upon completion, you'll be equipped with a fully functional job search automation tool as well as skills to build other AI agents.

Learning objectives
Create a functional AI planner that can automate searches and matching.
Develop custom tools for data parsing and web scraping sites.
Implement the think-plan-act agentic loop pattern through iterative code building.
Deploy a production-ready job search automation system with error handling and reliability features.
Master practical AI agent development techniques that transfer to other automation projects.

Concepts

Introduction

Building your first agent with GPT-OSS

Quick Start - Build Your First Agent

Demo - The complete agent in action
Environment setup - Ollama and GPT-OSS
How AI agents work - Intent, planning, tools, and reflection
Your first agent - Hello World with tools

Build Core Functionality

Exploring the code
Building `main.py`
Building the Location tool
Building the Weather tool
Building the Parks tool

Make It Smart - The Agentic Loop

Using an AI assistant to add error handling
Using an AI assistant to add documentation

Polish and Deploy

Adding memory
Reflection

Next Steps

Wrap-up and your next adventure