Future Skills · Learning hub

AI Agents Learning Hub

AI that plans, uses tools and finishes real tasks — and how to build one you can trust. 6 lessons, a Skill Check and guided projects — free.

Start with lesson 1 Take the Skill Check

Free tutorials
Practice tests
Projects

Lessons: 6
Reading time: 67 min
Skill Check: 12 questions
Projects: 2

About this skill

A chatbot answers your question and stops. An agent takes a goal — "find three internships that match my resume and draft an email for each" — breaks it into steps, uses tools like search, code and files, checks its own work and keeps going until the job is done. This track explains how agents actually work underneath: the reasoning loop, tool calling, memory and RAG. Then you build a small agent in Python and learn the guardrails that stop it from doing something expensive or embarrassing.

Who it is for: Students and freshers who already know the basics of AI and Python and want the most in-demand AI skill of 2026, plus developers who want to add agents to real products.

After this track you will be able to

Explain the difference between a chatbot, a workflow and an agent
Describe the agent loop: plan, act with a tool, observe, repeat
Define tools with a JSON schema and handle tool calls safely
Give an agent memory and grounded knowledge with RAG
Build a small working agent in Python
Add guardrails, human approval and evaluation before you ship

Not sure where to start? Take the AI Agents Skill Check 12 questions · about 10 minutes · instant level and a score for every topic Start

Learning path

Lessons, in order

Each lesson takes about 10 minutes and ends with a small task. 0 of 6 completed.

1
How Agents WorkWhat makes something an agent, and the plan–act–observe loop behind every one.
- What Are AI Agents? Chatbots vs Workflows vs Agents ExplainedAn agent is an AI system that is given a goal instead of a single question, and decides its own steps to reach it. Here is what that really means, how it differs from a chatbot, and when you should not use one. Beginner 9 min
- The Agent Loop: Planning, ReAct and Knowing When to StopEvery agent runs the same cycle: think about the next step, act with a tool, observe the result, repeat. See the loop step by step with a real example, and learn why stopping conditions matter as much as intelligence. Beginner 10 min
2
Tools, Memory & RAGHow agents call tools, remember context and answer from your own documents.
- Tool Calling Explained: How AI Agents Use Functions and APIsTool calling (also called function calling) lets a model ask your code to run a function with specific arguments. Learn how tools are described with JSON schemas, what a tool call looks like, how to write good tools, and where MCP fits. Intermediate 11 min
- Agent Memory and RAG: Giving AI Agents Context and KnowledgeModels forget everything between calls and know nothing about your private documents. Learn the types of agent memory, how retrieval-augmented generation (RAG) works with embeddings and vector search, and the mistakes that make RAG answers wrong. Intermediate 11 min
3
Building & Shipping AgentsWrite a small agent in Python, then make it safe, testable and affordable.
- Build Your First AI Agent in Python (No Framework Needed)Write a complete agent loop in about 60 lines of plain Python: tools, a step limit, error handling and a final answer. You can run it first with a scripted "fake model", then plug in any real LLM API. Intermediate 14 min
- AI Agent Guardrails, Safety and Evaluation: Ship Agents You Can TrustAn agent that can act can also act wrongly. Learn the guardrails real teams use — least privilege, human approval, prompt-injection defence, budgets and logging — and how to evaluate an agent before and after launch. Intermediate 12 min

Build

AI Agents projects

Finish the lessons, then prove the skill with a project you can show.

AI Agents project Build a Study Planner Agent with Tools A Python agent that reads your exam timetable and syllabus, calculates the days left and writes a realistic day-by-day revision plan — with a step limit, logs and tests. Tool callingAgent loopError handling Intermediate ~5 hours AI Agents project College Rules Assistant with RAG and Citations An assistant that answers questions about your college or company rules only from the official document, quotes the rule and says "I couldn’t find this" instead of guessing. RAGEmbeddingsPrompt design Intermediate ~6 hours

FAQ

AI Agents: common questions

Do I need to know machine learning to build agents?

No. Agent building is mostly software engineering: calling an AI model through an API, defining tools, handling errors and testing. You need comfortable Python and the ideas from the AI and Prompt Engineering tracks — not maths.

Which framework should I learn — LangChain, LlamaIndex, CrewAI or something else?

Learn the plain loop first, the way this track teaches it. Frameworks change every few months; the loop (model → tool call → result → model) does not. Once you understand it, any framework takes an afternoon to pick up, and you will know when you do not need one.

Will building agents cost me money?

You can learn almost everything for free. Most AI providers have free tiers or trial credits, and you can practise the loop with a small open model on your laptop. The projects here are designed to stay within free limits if you set a step limit, which we teach anyway.

Are AI agents safe to use at work?

Only with guardrails. An agent can send emails, delete files or spend money if you give it those tools. The last lesson covers permissions, human approval for risky actions, prompt-injection defence and logging — the things companies check before an agent goes live.