# How to Get Started with AI Development: A Complete Beginner's Guide Using Microsoft's Curriculum

> Start AI development with Microsoft's AI-For-Beginners curriculum. Clone the repo, set up your environment, and begin this comprehensive 12-week learning journey today.

- Repository: [Microsoft/AI-For-Beginners](https://github.com/microsoft/AI-For-Beginners)
- Tags: getting-started
- Published: 2026-07-03

---

**To start with AI development, clone the Microsoft AI-For-Beginners repository, set up the Conda environment defined in [`environment.yml`](https://github.com/microsoft/AI-For-Beginners/blob/main/environment.yml), and run the introductory script before launching Jupyter Lab to work through the 12-week curriculum.**

This guide walks you through how to get started with AI development using Microsoft's **AI-For-Beginners** repository, a comprehensive open-source curriculum that provides 24 lessons covering everything from symbolic AI to deep learning and ethics. The repository is structured as a modular educational platform with executable Jupyter notebooks, self-contained Python examples, and interactive quizzes designed to take you from zero to working AI implementations.

## Prerequisites and Repository Setup

The first step in your AI development journey involves obtaining the curriculum code and configuring a reproducible Python environment. The repository relies on Conda for dependency management and supports sparse checkout to minimize disk usage by excluding translation files.

### Clone the Repository with Sparse Checkout

To avoid downloading the entire repository including over 500MB of translations, use Git's sparse checkout feature. This command sequence clones only the essential English content and code:

```bash
git clone --filter=blob:none --sparse https://github.com/microsoft/AI-For-Beginners.git
cd AI-For-Beginners
git sparse-checkout set --no-cone '/*' '!translations' '!translated_images'

```

### Create the Conda Environment

The [`environment.yml`](https://github.com/microsoft/AI-For-Beginners/blob/main/environment.yml) file at the repository root defines a complete Python 3 environment with all necessary libraries, including Jupyter, NumPy, Matplotlib, OpenCV, and the full PyTorch stack (TorchVision, TorchText, and TorchData). Execute these commands to create and activate the environment:

```bash
conda env create -f environment.yml
conda activate ai4beg

```

## Verify Your Installation

Before diving into the curriculum, confirm your environment is correctly configured by running the minimal sanity-check script provided in the `examples/` directory.

Execute the following command from the repository root:

```bash
python examples/01-hello-ai-world.py

```

You should see the output:

```

Hello AI world!  🎉

```

This verification ensures that Python and the core dependencies are properly installed and accessible.

## Navigate the Curriculum Structure

The AI-For-Beginners repository is organized into a modular architecture that separates human-readable content from executable code. Understanding this structure helps you efficiently navigate the 12-week learning path.

The key directories include:

- **`lessons/`** – Contains 24 lessons organized by topic (symbolic AI, neural networks, computer vision, NLP, etc.), each with a [`README.md`](https://github.com/microsoft/AI-For-Beginners/blob/main/README.md) explaining concepts and linked Jupyter notebooks
- **`examples/`** – Self-contained Python scripts for quick experimentation without launching full notebooks
- **`etc/quiz-app/`** – A Vue.js application providing 50 interactive quizzes linked to specific lessons
- **`translations/`** – Language packs supporting over 40 languages (excluded if you used sparse checkout)

Each lesson follows a consistent pattern: pre-reading material in the [`README.md`](https://github.com/microsoft/AI-For-Beginners/blob/main/README.md), followed by implementation notebooks available in both **PyTorch** (e.g., `Perceptron.ipynb`) and **TensorFlow** (e.g., `Perceptron.ipynb` or `TensorFlow.ipynb`) variants, plus optional lab assignments.

## Run Your First AI Code

Once your environment is verified, launch Jupyter Lab to begin interactive development. This interface allows you to execute the curriculum notebooks step-by-step.

Start the server from the repository root:

```bash
jupyter lab

```

Navigate to [`lessons/1-Intro/README.md`](https://github.com/microsoft/AI-For-Beginners/blob/main/lessons/1-Intro/README.md) to begin with the course overview, or jump directly to `lessons/3-NeuralNetworks/03-Perceptron/Perceptron.ipynb` to implement your first neural network. Each notebook contains executable cells that implement the concepts discussed in the corresponding lesson.

For example, to load and visualize image data using PyTorch as demonstrated in the computer vision lessons, you would execute:

```python
import torch, torchvision, matplotlib.pyplot as plt
from torchvision import transforms, datasets

# Load a small sample dataset

transform = transforms.Compose([transforms.Resize((28, 28)),
                                transforms.ToTensor()])
mnist = datasets.MNIST(root='data', train=True, download=True, transform=transform)

# Simple data loader

loader = torch.utils.data.DataLoader(mnist, batch_size=64, shuffle=True)

# Visualise one batch

images, labels = next(iter(loader))
plt.imshow(images[0][0], cmap='gray')
plt.title(f'Label: {labels[0].item()}')
plt.show()

```

## Interactive Learning Tools

Reinforce your understanding by running the integrated quiz application. This Vue.js tool provides knowledge checks linked to specific lessons.

From the repository root, execute:

```bash
cd etc/quiz-app
npm install
npm run serve

```

Then open `http://localhost:8080` in your browser to access the interactive quizzes. The application draws from the same curriculum structure as the notebooks, ensuring alignment with your current learning module.

## Summary

- **Clone efficiently** using sparse checkout to save disk space by excluding the `translations/` directory while keeping all code and lessons.
- **Install dependencies** via `conda env create -f environment.yml` to obtain the exact Python environment used by the curriculum maintainers, including PyTorch and TensorFlow support.
- **Verify setup** by running `python examples/01-hello-ai-world.py` before attempting complex notebooks.
- **Follow the modular path** starting from `lessons/1-Intro/` and progress through the 12-week structure, choosing either PyTorch or TensorFlow implementations based on your preference.
- **Test knowledge** using the Vue.js quiz app in `etc/quiz-app/` after completing lesson sections.

## Frequently Asked Questions

### Do I need to know Python before starting this curriculum?

Yes, basic Python knowledge is assumed. The [`environment.yml`](https://github.com/microsoft/AI-For-Beginners/blob/main/environment.yml) file specifies Python 3, and the notebooks in `lessons/` use standard Python syntax with scientific computing libraries. While the curriculum teaches AI concepts from scratch, you should understand variables, functions, and basic data structures before diving into the neural network implementations in `lessons/3-NeuralNetworks/`.

### Can I use TensorFlow instead of PyTorch?

Yes. The repository provides parallel implementations for most lessons. Each neural network and computer vision lesson contains both `*PyTorch.ipynb` and `*TensorFlow.ipynb` files (or `*TF.ipynb` variants). The [`environment.yml`](https://github.com/microsoft/AI-For-Beginners/blob/main/environment.yml) includes dependencies for both frameworks, so you can choose your preferred library when opening the corresponding notebook in Jupyter Lab.

### How long does it take to complete the entire curriculum?

The curriculum is designed as a 12-week program with 24 lessons, implying roughly two lessons per week. However, the self-paced nature of the repository allows you to adjust timing based on your schedule. Each lesson includes pre-reading, a notebook exercise (1-2 hours), and optional labs, making the full course approximately 24-40 hours of hands-on work.

### What hardware requirements are needed for the practical exercises?

Most exercises in the beginner lessons run on standard CPUs with 8GB RAM. The [`environment.yml`](https://github.com/microsoft/AI-For-Beginners/blob/main/environment.yml) installs CPU-only versions of PyTorch by default. However, as you progress to deeper neural networks in `lessons/4-ComputerVision/` and `lessons/5-NLP/`, having a CUDA-compatible GPU significantly accelerates training times for the convolutional and transformer models, though it is not strictly required to complete the curriculum.