In this hands-on course, you will learn how to build AI agents with Python and Google ADK, connect agents to tools and memory, work with RAG and MCP, and build multi-agent applications from the ground up. You will start with the fundamentals of AI agents and learn how the agent loop works, how tools are used, and how different agent patterns can solve different types of problems. You will then build five important patterns: sequential, hierarchical, routing, multi-agent, and autonomous agents. From there, you will work with Google ADK and build a practical StudyBuddy application. You will learn how to give an agent tools, add session memory, debug agent behavior, and understand how an agent SDK can simplify development.
The course also covers prompt engineering and context engineering, including few-shot prompting, structured outputs, prompt injection concepts, and techniques for providing better context to agents. You will then learn LangChain fundamentals and move into Retrieval-Augmented Generation (RAG). You will understand embeddings, similarity search, vector databases, chunking, retrievers, and how to build a complete RAG application using your own documents. Next, you will learn Model Context Protocol (MCP), including MCP clients and servers, the handshake process, tools, and how to connect an MCP server to an AI agent.
You will also learn how to deploy AI applications, work with secrets, add basic guardrails, and troubleshoot common problems during development and deployment. Finally, you will build ResolveAI, the course’s flagship capstone project. This multi-agent application is developed using a spec-driven workflow and deployed to Google Cloud, giving you an opportunity to bring together concepts from across the course. The course also includes quizzes, complete source code, practical exercises, and an Interview Mastery module with AI-agent interview questions and answers.
