Everything, in the right order.
8 modules · 41 lessons · 10 hours. Start at the top, or jump to what you need.
Your progress
Loading your progress…
Start the course
What Is Agentic AI?
Foundations: LLMs & the Agentic Shift
Build correct mental models for how large language models work and why wrapping them in a loop with tools and memory creates something qualitatively new: an agent.
What Is Agentic AI?
The shift from tools that answer to systems that act
12 minHow LLMs Actually Work
Tokens, transformers, and next-token prediction — the intuitive version
14 minContext Windows & Prompting
The model's working memory — and how to fill it well
13 minCapabilities & Limitations
What LLMs are brilliant and terrible at — and why agents exist
12 minThe Agent Loop: A First Look
How a simple loop turns a text predictor into an actor
11 minCore Agent Concepts
The reasoning loop, tool use and function calling, memory, planning, and reflection — the universal building blocks underneath every agent framework.
Anatomy of an Agent
The four parts under every framework
13 minReasoning & the ReAct Pattern
Interleaving thought and action
14 minTool Use & Function Calling
How an agent reaches outside its own head
15 minMemory: Short-Term to Long-Term
Helping agents remember beyond a single context window
15 minPlanning & Task Decomposition
Turning a big goal into doable steps
13 minBuilding Agents: Hands-On
Write agents from scratch, then with frameworks. Master tool design, retrieval-augmented generation, structured outputs, and the engineering that makes agents dependable.
Your First Agent From Scratch
A complete agent loop in ~40 lines, no framework
16 minDesigning Tools Agents Can Use
Tool design is prompt engineering for actions
14 minRetrieval-Augmented Generation for Agents
Giving agents access to your knowledge
16 minStructured Outputs & Reliability
Making model output safe to program against
13 minThe Framework Landscape
LangGraph, CrewAI, AutoGen, Agents SDKs — what to use when
15 minBuilding with a Graph Framework
Stateful, controllable agents with LangGraph
17 minMulti-Agent Systems
Orchestration patterns, agent-to-agent communication, the Model Context Protocol, roles and debate, and when multiple agents beat one.
Why (and When) Multi-Agent
More agents isn't automatically better
12 minOrchestration Patterns
Supervisors, hierarchies, pipelines, and swarms
15 minHow Agents Communicate
Messages, shared state, and handoffs
13 minMCP & A2A: Interop Protocols
The USB-C of AI tools and the language of agent-to-agent
16 minRoles, Debate & Collaboration
Getting better answers from disagreement
13 minAdvanced Reasoning & Planning
Chain-of-thought, ReAct, reflection, tree- and graph-based planning, and the techniques that let agents tackle long-horizon, multi-step problems.
Chain-of-Thought & Reasoning Models
Letting models think before answering
14 minReflection & Self-Correction
Agents that check their own work
13 minTree- & Graph-of-Thought Planning
Searching over many possible reasoning paths
13 minLong-Horizon & Stateful Tasks
Staying coherent over hundreds of steps
14 minProduction Engineering for Agents
Evaluation, observability and tracing, cost and latency optimization, guardrails, caching, and deploying agents that survive contact with real users.
Evaluating Agents
If you can't measure it, you can't ship it
16 minObservability & Tracing
Seeing inside an agent's decisions
13 minCost & Latency Optimization
Fast, cheap, good — engineering the trade-offs
14 minReliability & Guardrails
Designing for the 20% that goes wrong
14 minDeploying & Scaling Agents
From notebook to always-on service
14 minContext Engineering
The successor discipline to prompt engineering
15 minSafety, Security & Governance
Prompt injection and the agent attack surface, permissions and sandboxing, alignment and human oversight, plus the governance and policy landscape.
Prompt Injection & the Attack Surface
The defining security problem of agents
15 minSecuring Agentic Systems
Least privilege for software that acts on its own
14 minAlignment & Human Oversight
Keeping capable agents under control
13 minResponsible AI & Governance
Policy, risk, and doing this responsibly
13 minAgent Failure Modes & Red-Teaming
How agents break, and how to find out first
13 minFrontiers & Capstone
Computer-use and coding agents, agentic RAG and memory research, the open problems, and a capstone that ties everything together into a portfolio project.
Computer-Use & Browser Agents
Agents that click, type, and browse like a person
14 minCoding Agents
The killer app of agentic AI
15 minThe Memory & Learning Frontier
Agents that get better over time
13 minOpen Problems in Agentic AI
The hard questions the field is still answering
12 minCapstone: Build a Production Agent
Put it all together into a portfolio project
18 min