AI Agents Crash Course: Build with Python & OpenAI

Learn to build agentic AI solutions with Python, OpenAI SDK, AgentBuilder, tools, RAG, and guardrails - in just 4 hours!

What you'll learn
Build functional AI agents using Python and the OpenAI SDK
Implement tool calling, memory, and streaming responses in your agents
Use prompt engineering and context engineering to control agent behavior
Integrate Retrieval-Augmented Generation (RAG) using embedding databases
Enforce safety with guardrails and prompt adherence techniques
Orchestrate multi-agent systems with task decomposition and hand-offs
Deploy AI agents to the cloud with authentication and secure setup
Trace and debug agent behavior using OpenAI’s built-in tools

Requirements
Basic experience with Python programming
An OpenAI account with API access
A GitHub account to use GitHub Codespaces for development
Familiarity with basic coding tools (e.g., running scripts, editing code)

Description
Building intelligent AI agents can feel overwhelming. Between OpenAI’s complex SDK, retrieval-augmented generation (RAG), tool-calling, memory, and prompt engineering, it’s hard to know where to start.This crash course is your shortcut: in just a few hours, you'll go from zero to deploying your own functional, real-world agentic AI system. You'll go hands-on building agentic AI systems with Python, and also visually in the no-code AgentBuilder environment.You’ll build a smart nutrition assistant that:Uses OpenAI’s Agents SDK and AgentKit to understand and respond to promptsCalls external tools and APIsLeverages memory and RAG for contextual intelligenceIncludes guardrails to behave safely and reliablyCan be deployed to the cloud with authenticationWhether you're a developer, data scientist, or AI-curious engineer, this hands-on course gives you a complete end-to-end agentic AI foundation – without getting buried in theory or outdated code.What You’ll LearnHow to build AI agents with Python + OpenAI’s Agents SDKVisually developing and deploying agentic systems with AgentKit, AgentBuilder, ChatKit, and EvalsTool calling, streaming, and tracing techniquesBest practices in prompt engineering and context designHow to integrate memory and RAG for deeper contextual reasoningDeploying your agent securely with authentication and guardrailsHow to build multi-agent systems with task delegation and parallel executionWho This Course is ForEngineers and developers with basic Python experienceAI/ML professionals looking to quickly learn agent orchestrationProduct builders and technical leads exploring agentic workflowsLearners who want to build, not just read about agentsAbout the InstructorsYour instructors combine deep industry experience with a passion for clear, actionable teaching.Frank Kane spent 9 years at Amazon and IMDb, where he built large-scale recommender systems and led engineering teams. He holds 17 patents in machine learning and distributed systems and has taught over 1 million students through his company, Sundog Education.Zoltan C. Toth brings over two decades of experience in AI infrastructure and data systems. As a former principal instructor and Solutions Architect Databricks and Data Engineering lead at startups, he’s helped companies around the world scale their analytics and AI platforms. Zoltan also teaches AI and data engineering at the Central European University.Together, Frank and Zoltan guide you step-by-step through building agents the right way: with real code, real tools, and production-ready techniques.Ready to build your first AI agent, fast?Enroll now and start building today.

Who this course is for:
Learners who prefer a project-based, fast-paced crash course over lengthy theory, Python developers who want to quickly get hands-on with building AI agents, Software engineers and technical leads looking to integrate AI agents into real-world products, Entrepreneurs, indie hackers, and product builders interested in agentic AI workflows, Anyone curious about how to use the OpenAI SDK to build smart, autonomous applications