₹0 for students · Certificate · Beginner

AI for Beginners

24 lessons, from perceptrons to large language models

Start free in the app

24 lessons · ~27 hours · 30-question final exam · pass mark 70%

About the course

What you’ll learn

A practical, beginner-friendly tour of artificial intelligence: symbolic AI, neural networks, computer vision, natural language processing, reinforcement learning, multi-agent systems and responsible AI. Each lesson has a short study guide, links to the full lesson, notebooks and labs, and a quiz. Pass all 24 lesson quizzes, then the final exam, to earn a verifiable certificate.

Neural networksComputer visionNLPPyTorchTensorFlowTransformersReinforcement learningResponsible AI

Part of the career tracks: Machine Learning Engineer, GenAI and LLM Builder

How the certificate works

  1. Read each lesson’s study guide, then the full lesson and any notebooks or exercises.
  2. Pass every lesson quiz with 70% or more.
  3. Pass the final exam: 30 questions in 45 minutes.
  4. Get a certificate with a public verification link, and add it to LinkedIn.

Syllabus

25 lessons in 8 parts

Introduction to AI

1
Introduction and History of AI
45 min · AI targets tasks we can't describe as an explicit algorithm, like estimating age from a photo.

Symbolic AI

2
Knowledge Representation and Expert Systems
60 min · DIKW: data → information → knowledge → wisdom.

Introduction to Neural Networks

3
Perceptron
60 min · Rosenblatt's Mark-1 (1957) was a hardware perceptron for recognising shapes.
4
Multi-Layered Perceptron and Building Our Own Framework
75 min · Parameters θ = weights and biases; training minimises the loss by changing them.
5
Intro to Frameworks (PyTorch/TensorFlow) and Overfitting
90 min · Frameworks provide tensor operations (CPU/GPU) and automatic gradients.

Computer Vision

6
Intro to Computer Vision with OpenCV
60 min · Images are NumPy arrays: H×W (grayscale) or H×W×3 (colour).
7
Convolutional Neural Networks and CNN Architectures
90 min · A convolutional filter slides a small kernel over the image to detect a pattern.
8
Pre-trained Networks and Transfer Learning
75 min · Transfer learning reuses a model pre-trained on a big dataset (e.g. ImageNet).
9
Autoencoders and Variational Autoencoders
75 min · Autoencoder = encoder → small latent vector → decoder that rebuilds the input.
10
Generative Adversarial Networks and Artistic Style Transfer
75 min · Generator: random vector → image. Discriminator: image → real or fake.
11
Object Detection
75 min · Detection = class + bounding box for each object.
12
Semantic Segmentation and U-Net
60 min · Segmentation = classifying every pixel of an image.

Natural Language Processing

13
Representing Text: Bag-of-Words and TF-IDF
60 min · Text → tokens (characters, words, sub-words) → numbers via a vocabulary → one-hot vectors.
14
Semantic Word Embeddings: Word2Vec and GloVe
60 min · Embeddings are dense, low-dimensional word vectors; one-hot vectors are huge and unrelated.
15
Language Modeling: Training Your Own Embeddings
60 min · Language models learn from unlabelled text (self-supervised), which is abundant.
16
Recurrent Neural Networks
90 min · RNNs process a sequence step by step, passing a state from step to step.
17
Generative Recurrent Networks
75 min · RNN patterns: one-to-one, one-to-many, many-to-one, many-to-many (seq2seq).
18
Attention Mechanisms, Transformers and BERT
90 min · Attention weights all encoder states when producing each output token.
19
Named Entity Recognition
60 min · NER finds entities (people, places, dates, chemicals, diseases…) in text.
20
Large Language Models and Prompt Engineering
60 min · GPT = Generative Pre-trained Transformer; it can solve tasks without task-specific training.

Other AI Techniques

21
Genetic Algorithms
45 min · Genetic algorithms evolve a population of solutions (John Holland, 1975).
22
Deep Reinforcement Learning
90 min · RL learns by doing: environment + reward function; no labelled answers.
23
Multi-Agent Systems
60 min · Emergence: simple agents together can produce complex behaviour.

AI Ethics

24
Ethical and Responsible AI
45 min · Today's AI is mathematical pattern-finding: a powerful tool that can be misused.

Bonus

25
Bonus: Multi-Modal Networks (CLIP, VQGAN, DALL·E)Optional
45 min · CLIP matches images and text using contrastive training on image–caption pairs.

Course content is adapted from Microsoft's "AI for Beginners" curriculum (https://github.com/microsoft/AI-For-Beginners), used under the MIT License, Copyright (c) Microsoft Corporation. Study guides and quiz questions are written by Vikosha. Certificates are issued by Vikosha and are not issued or endorsed by Microsoft. Licence: MIT License, Copyright (c) Microsoft Corporation. View the original curriculum