Free Learning Resources

AI Models & Agents Learning Roadmap

From beginner to advanced with hands-on courses and developer resources

Getting Started
Last updated: Aug 29, 2025

Scope: LLMs → agents → MCP (Model Context Protocol) → Google Agent-to-Agent (A2A) → evals & safety.

All links are free to access (some may require free sign-up).

Quick Map of Core Concepts

LLM (Large Language Model)

Transformer-based next-token predictor; gives you language understanding/generation. See a clear visual intro by 3Blue1Brown.

Agent

An LLM wrapped with tools, memory, planning, and feedback loops so it can act, not just answer.

MCP (Model Context Protocol)

An open protocol that standardizes how AI apps connect to tools and data (think “USB-C for tools & context”).

A2A (Agent-to-Agent)

An open protocol so heterogeneous agents can discover, negotiate, and collaborate across orgs, stacks, and clouds.

Beginner Track
Start here — fundamental concepts and introductions
1

Transformers, the tech behind LLMs

Video

Visual, intuition-first walkthrough of attention & Transformers by 3Blue1Brown.

Watch on YouTube
2

The Illustrated Transformer

Article

Friendly diagrams explaining attention, encoder/decoder, and how the original Transformer works.

Read Article
3

Stanford CS25 — Transformers United

Course

Survey talks on Transformer applications across domains; great for context and inspiration.

View Course
4

Hugging Face LLM Course (Chapter 1)

Interactive

Foundations of Transformers and the Hugging Face ecosystem; run examples, fine-tune, share on the Hub.

Start Learning
5

Model Context Protocol — Introduction

Docs

What MCP is, why it exists, and the core concepts (servers, clients, resources, tools).

Read Introduction
6

Google Agent-to-Agent (A2A) — Overview

Blog

What A2A solves (agent discovery, negotiation, secure collaboration) and where it fits in Google's agent stack.

Read Blog Post
Intermediate Track
Build, connect, and evaluate — hands-on projects and frameworks

Hugging Face AI Agents Course

Hands-on

Concepts + frameworks (LangGraph, LlamaIndex, smolagents) and a capstone where you ship an agent.

Start Course

Hugging Face MCP Course

Hands-on

Build MCP servers/clients; then compose them into an end-to-end app.

Start Course

Building Systems with ChatGPT API

DeepLearning.AI

Patterns for tool-calling, routing, multi-step workflows.

Take Course

Evaluating AI Agents

DeepLearning.AI

How to trace, test, and iterate on agent performance with structured evals.

Take Course

LlamaIndex — Starter Tutorial

RAG

Minimal app, then add RAG; clear examples for agents + retrieval pipelines.

View Tutorial

Haystack — First RAG Pipeline

RAG

Build a retrieval-augmented QA pipeline; great for seeing the “plumbing.”

LangChain — Build an Agent

Tools

Tool-calling agents with clear, up-to-date examples.

View Tutorial

A2A Protocol — Purchasing Concierge

Codelab

Deploy multiple agents on Cloud Run/Agent Engine and watch them collaborate via A2A.

Try Codelab
Advanced Track
Theory, surveys, and specifications — deep dive into research

Research Papers

Survey: LLM-based Autonomous Agents

arXiv

Architecture patterns (planning, tools, memory, reflection) and applications.

Read Paper

Survey: LLM-based Multi-Agent Systems

arXiv

Collaboration mechanisms and environments; recent advances & open problems.

Read Paper

Academic Content

Stanford CS224N — Reasoning & Agents

Video

Modern view on reasoning in LLMs and implications for agent design.

Watch Lecture

Berkeley MOOC: LLM Agents

Course

Full agentic computing syllabus; lectures are public.

Protocol Specifications

MCP — Official Spec & SDKs

Protocol

Canonical spec + official Python/TypeScript SDKs.

A2A — Protocol & Repos

Protocol

Open protocol for inter-agent collaboration; docs, SDKs, and samples.

Expert Insights

Understanding Reasoning LLMs

Blog

Training and inference-time strategies for stronger reasoning by Sebastian Raschka.

Read Article
Developer Track
Code-first resources, tools, and frameworks for hands-on development

AI Coding Tools

Claude Code — Anthropic's CLI for Claude
Cursor — AI-powered code editor
Amazon Q Developer — AWS AI coding assistant
GitHub Copilot — AI pair programmer

Protocols & Runtimes

MCP (Model Context Protocol)
A2A (Agent-to-Agent)

LangGraph & Orchestration

LangGraph Official — Graph-based workflows
LangGraph Academy — Official course
LangGraph Tutorials — Documentation
Sam Witteveen's Tutorials — 33+ notebooks

RAG & SDKs

Amazon Bedrock — Managed foundation models
OpenAI Cookbook — Patterns & examples
LlamaIndex — RAG framework
Haystack — Search & QA

Evaluation & Safety

Bedrock Guardrails — AWS safety controls
promptfoo — Testing CLI/SDK
DeepEval — Unit-test-style evals
NeMo Guardrails — Safety rails
Guardrails AI — I/O validators

AWS Resources

Bedrock AgentCore — Suite of Agent Services
Strands Agents SDK — Agent SDK
Getting Started Guide
Suggested 2–4 week learning path to master AI agents
Week 1

Watch the 3Blue1Brown video and skim Illustrated Transformer to cement intuition.

Week 2

Do HF LLM Course (Chapter 1), then jump into HF Agents and HF MCP for hands-on work.

Week 3

Build a tiny RAG app (LlamaIndex or Haystack), then wrap it with LangGraph.

Week 4

Explore MCP for tools/data and the A2A codelab to see cross-agent collaboration.

Deploy

Add evals and guardrails before you scale/deploy.

💡 Pro Tip

Everything above is free to access. Some platforms (Hugging Face, DeepLearning.AI, Google Cloud) may ask you to create a free account to save progress or run hosted examples.