AI DevelopmentIntermediate · 8 weeks

Building AI Agents

Go beyond chatbots. Build AI systems that plan, act, and get things done.

AI agents are the next frontier — systems that don't just answer questions but take actions, use tools, plan multi-step tasks, and operate with increasing autonomy. This programme teaches you to build real AI agents from first principles, progressing from simple tool-using assistants to multi-agent systems capable of complex, autonomous workflows.

Self-PacedProject-Based

Prerequisites

—Basic Python programming knowledge
—Familiarity with at least one AI assistant
—Understanding of APIs and JSON data formats

What you'll be able to do

✓Understand the architecture of AI agent systems — tools, memory, planning, and execution
✓Build function-calling and tool-using agents from scratch
✓Implement memory systems: short-term, long-term, and episodic memory
✓Design and build multi-step planning agents for complex tasks
✓Work with major agent frameworks: LangChain, LangGraph, and AutoGen
✓Build and deploy a complete multi-agent system as a capstone project
✓Apply responsible AI principles to autonomous agent design

Curriculum

1. Introduction to AI Agents

·What is an AI agent and why it matters now
·Agents vs. chatbots vs. autonomous systems
·The agent loop: perceive, plan, act, observe
·Real-world examples of AI agents in production

2. Agent Architecture

·The ReAct pattern: Reason + Act
·Function calling and tool use in LLMs
·Structured outputs and schema enforcement
·Designing robust agent architectures

3. Tools

·Giving agents access to external tools
·Building custom tools for your agent
·Web search, code execution, and file system tools
·Tool selection and routing strategies

4. Memory

·Short-term memory: conversation history management
·Long-term memory: vector stores and retrieval
·Episodic memory: learning from past agent runs
·Memory compression and summarisation strategies

5. Context and Planning

·Context window management for long-running agents
·Task decomposition and sub-task planning
·Sequential vs. parallel task execution
·Handling ambiguity and uncertainty in agent tasks

6. Agent Frameworks

·LangChain: the foundational agent framework
·LangGraph: stateful, multi-step agent workflows
·AutoGen (Microsoft): multi-agent conversation patterns
·CrewAI: role-based multi-agent systems
·Choosing the right framework for your use case

7. Multi-Agent Systems

·Why single agents have limits
·Orchestrator and specialist agent patterns
·Agent communication and handoffs
·Debugging and observing multi-agent systems

8. Practical Agent Projects

·Project 1: Research agent with web search and synthesis
·Project 2: Code generation and testing agent
·Project 3: Autonomous data analysis agent
·Project 4 (Capstone): Complete multi-agent workflow

Format & Duration

Duration

8 weeks

Format

Self-Paced, Project-Based

Level

Intermediate