Amazon AgentCore: Scale Your Agentic AI to Production

Tired of building AI prototypes that never make it to production?
You’re not alone. Many engineers can build impressive agentic AI demos—but hitting a wall when trying to scale those systems into production is common.

This course solves that problem.

You’ll learn how to use Amazon Bedrock AgentCore—part of AWS’s cutting-edge generative AI stack—to deploy real, secure, scalable agent systems. You’ll take a working OpenAI Agents SDK project and transform it into a production-grade service, using AgentCore’s built-in memory, identity, tools, and observability features.

By the end of this course, you won’t just understand agentic AI—you’ll have deployed one.

What You’ll Learn

• How to use Amazon AgentCore to host your AI agents serverlessly in production

• Add memory to your agents (short-term and long-term)

• Handle user identity and secure authentication in agent workflows

• Integrate real tools, APIs, and third-party data using Bedrock’s Gateways

• Monitor and debug agents using AgentCore’s observability features

• Build trust in your agent with Policies and Evaluators

• Build a complete hands-on agentic AI project using the OpenAI Agents SDK

Why Amazon AgentCore?

Amazon Bedrock AgentCore provides a serverless runtime purpose-built for agentic AI. It handles scaling, security, and tool integrations so you don’t have to. With first-class support in AWS, it’s the fastest way to take your generative AI project from experiment to enterprise.

Who This Course Is For

• AI engineers and developers who’ve built agent prototypes—but haven’t shipped them

• ML practitioners ready to operationalize generative AI

• Software engineers looking to upskill in AWS AI tools and infrastructure

• Builders who want hands-on, project-based experience with agent systems in production

If you’ve been exploring agentic AI or the OpenAI Agents SDK, this course will show you how to make it real—on a secure, scalable production stack.

About the Instructor

Hi, I’m Frank Kane. I spent 9 years at Amazon and IMDb, where I helped build and lead the AI systems behind some of the most-visited websites on the planet.

Since leaving Amazon, I’ve taught over one million students around the world how to succeed in machine learning and data science through Sundog Education.

This course brings my real-world engineering experience at Amazon together with today’s most powerful agentic AI tools—so you can stop prototyping and start deploying.

What You'll Walk Away With

By the end, you’ll have:

• A working, full-featured agentic AI system deployed with Amazon AgentCore

• The confidence to scale, monitor, and maintain your own production agents

• Practical experience that applies directly to your work or portfolio

Please Note:

Following along hands-on with the project in this course requires an OpenAI developer account and an AWS account, as well as a Python development environment. Total costs should not exceed a few dollars, or you can just watch the videos without incurring any cloud costs.