Learning hub

Learn agentic AI

Free, practical guides for students and engineers who want to understand, build, and ship AI agents. Start from first principles and work your way to production-grade autonomous systems — no fluff, no paywall.

  • 21 in-depth guides
  • Beginner to advanced
  • Updated 2026

This is the learning hub for everyone who wants to learn agentic AI — the field of building software that reasons about a goal, takes real actions with tools, and keeps going until the work is done. Each guide is written to be genuinely useful: clear definitions, concrete examples, real concepts like ReAct, tool calling, RAG, vector stores, and orchestration — and zero filler.

The curriculum moves in a deliberate order. You'll start with what an AI agent actually is and how it differs from the generative AI you already know. From there you'll compare frameworks, build your first agent, give it tools and memory, and finally orchestrate multi-agent systems ready for production.

Who it's for: students exploring the topic for a course or project, software engineers adding agents to a product, and technical leaders who need an accurate mental model before committing a roadmap. If you can read code and call an API, you have everything you need to begin.

The curriculum

Every agentic AI guide, in one place

Twenty-one focused lessons covering the full stack of building AI agents — from fundamentals to RAG, evaluation, security, and deployment. Read them in order, or jump straight to what you need.

What is agentic AI?

Start here

The fundamentals: agency, the perceive–reason–act–observe loop, and the five components of every agent.

Learn more

Agentic AI vs generative AI

Why generative AI answers and agentic AI acts — and how one contains the other. Clear up the confusion fast.

Learn more

LLM agents explained

How large language models become agents through reasoning, tool calling, and the ReAct pattern.

Learn more

How to build AI agents

A practical, step-by-step build: from a single prompt to a working autonomous agent that uses tools.

Learn more

AI agent frameworks

Compare the leading frameworks and SDKs so you can choose the right stack for your project.

Learn more

AI agent tools

Function calling, APIs, and the tool patterns that let an agent take real actions in the world.

Learn more

AI agent memory

Short-term context vs. long-term memory, vector stores, and retrieval-augmented generation (RAG).

Learn more

Agentic workflows

Patterns like planning, reflection, routing, and evaluator–optimizer that make agents reliable.

Learn more

Multi-agent systems

Orchestrator–worker designs, delegation, and coordination for teams of specialist agents.

Learn more

RAG for AI agents

Retrieval-augmented generation: ground agent answers in your real, current data with retrieval.

Learn more

Function calling

How LLMs emit structured tool calls from a JSON schema — the mechanism beneath every tool-using agent.

Learn more

Vector databases

Embeddings, similarity search, and ANN indexes — the storage layer behind RAG and long-term memory.

Learn more

AI agent architecture

The canonical components — model, planner, tools, memory, orchestration, guardrails — and how they fit.

Learn more

AI agent orchestration

Routing, handoffs, and orchestrator–worker patterns for coordinating steps and multiple agents.

Learn more

Autonomous agents

What 'autonomous' really means, the levels of autonomy, and how to keep self-directed agents safe.

Learn more

AI agents vs chatbots

Pursuing goals with tools and actions versus answering one turn at a time — and when each is enough.

Learn more

Prompt engineering

System prompts, tool descriptions, few-shot examples, and context engineering for reliable agents.

Learn more

AI agent evaluation

Measure task success, tool accuracy, and grounding — and run agent evals in CI to catch regressions.

Learn more

AI agent observability

Traces, spans, token and cost signals — see exactly what an agent did, and debug it with confidence.

Learn more

Deploying AI agents

Take agents to production: hosting, state, concurrency, rollouts, and human-in-the-loop checkpoints.

Learn more

AI agent security

Prompt injection, least privilege, sandboxing, and the layered guardrails that keep agents safe.

Learn more
The learning path

A six-step path from zero to production

Follow these stages in order. Each one maps to a guide and builds directly on the last.

  1. Understand the fundamentals

    Learn what agentic AI is, how the agent loop works, and how it differs from generative AI. Build the mental model before writing any code. → 'What is agentic AI?'

  2. Compare frameworks

    Survey the agent frameworks and SDKs, weigh their trade-offs, and pick the stack that fits your language and use case. → 'AI agent frameworks'

  3. Build your first agent

    Wire an LLM into a reasoning loop and ship a minimal working agent that completes a real task end to end. → 'How to build AI agents'

  4. Add tools & memory

    Give the agent function-calling tools to act and a memory layer (with a vector store and RAG) so it remembers context. → 'AI agent tools' + 'AI agent memory'

  5. Orchestrate multi-agent teams

    Decompose hard goals across specialist agents with an orchestrator, and apply workflow patterns for reliability. → 'Multi-agent systems' + 'Agentic workflows'

  6. Ship to production

    Add evaluation, guardrails, observability, and SOC 2-aligned security, then deploy with confidence. → Explore the docs and SDKs.

21

Free guides

no paywall

6

Path stages

zero to prod

1 day

First prototype

is realistic

100%

Hands-on

concepts + code

How to use this path

New to the topic? Go top to bottom. Already building? Jump to the stage you're stuck on — each guide stands on its own, with cross-links back to the fundamentals when you need them.

Explore by topic

Core agentic AI concepts you'll master

The vocabulary of the field — every term below is covered across the guides above.

FAQ

Learning agentic AI, answered

Start with the fundamentals: read 'What is agentic AI?' to understand the perceive–reason–act–observe loop, then compare it with generative AI so the boundary is clear. Once the mental model clicks, move into the hands-on guides — building your first agent, wiring up tools, and adding memory. Every guide here is free, self-contained, and ordered so you can follow it top to bottom without prerequisites piling up.

Get started

Ready to build your first AI agent?

You've learned the theory — now ship the real thing. Start free, no credit card required, and go from guide to running agent in minutes.