AI Foundations: From Data to Large Language Models | Nesqual TechAll coursesAI Foundations: From Data to Large Language Models
Understand how modern AI works, from the maths and data to neural networks, LLMs and AI security.
- Access
- Free
- Level
- Beginner
- Duration
- 41 h
- Sections
- 84
Start free Course outline
What artificial intelligence is
- A working definition of AILesson25 min
- AI, machine learning, deep learning and generative AILesson25 min
- A short, honest history of AILesson30 min
- What today's AI can and cannot doLesson25 min
- Models, programs and the words we will useLesson30 min
- Chapter quiz: what artificial intelligence isQuiz15 min
The maths you actually need
- Numbers as vectorsLesson30 min
Matrices as tables and transformations
Lesson30 min
Functions, slopes and the chain ruleLesson35 minProbability and Bayes' ruleLesson35 minStatistics for dataLesson35 minChapter quiz: the maths you actually needQuiz15 minPython for AI
- Why Python, and values, variables and typesLesson30 min
- Lists, dictionaries and loopsLesson30 min
- Functions and modulesLesson30 min
- Reading and writing files and CSVLesson30 min
- Vectors and matrices in plain PythonLesson35 min
- Chapter quiz: Python for AIQuiz15 min
- Lab: statistics and vectors in plain PythonLab40 min
Working with data
- Where data comes from and what a dataset isLesson25 min
- Data quality problems and cleaningLesson30 min
- Turning the world into numbersLesson30 min
- Train, validation and test splits, and data leakageLesson35 min
- Bias in data, labelling and the cost of ground truthLesson30 min
- Working with data quizQuiz15 min
Classical machine learning
- Supervised, unsupervised and reinforcement learningLesson25 min
- Linear regression worked by handLesson30 min
- Classification with logistic regression and decision boundariesLesson30 min
- Decision trees, random forests and k-nearest neighboursLesson35 min
- Clustering with k-means, step by stepLesson30 min
- Evaluating models and choosing one for a problemLesson35 min
- Classical machine learning quizQuiz15 min
- Lab: k-nearest neighbours from scratchLab40 min
How training really works
- Loss functions, a score for being wrongLesson30 min
- Gradient descent, walking downhill step by stepLesson35 min
- Learning rate, batches, epochs and stochastic gradient descentLesson30 min
- Underfitting, overfitting and how to fight themLesson35 min
- Hyperparameters, cross-validation and reading training curvesLesson35 min
- How training really works quizQuiz15 min
- Lab: gradient descent from scratchLab40 min
Neural networks
- The artificial neuron, a weighted voteLesson25 min
- Why one neuron is not enough, the XOR problemLesson25 min
- Activation functions and why non-linearity mattersLesson25 min
- The forward pass, computed by handLesson30 min
- Backpropagation, assigning the blameLesson35 min
- Training networks in practiceLesson30 min
- Lab: build a neural network that learns XORLab40 min
- Neural networks quizQuiz15 min
Deep learning architectures
- Why architecture matters, and convolutional networksLesson30 min
- Sequence models, RNNs and LSTMsLesson30 min
- Embeddings, meaning as geometryLesson30 min
- Attention, explained from scratchLesson35 min
- The transformer, block by blockLesson35 min
- Generative models for images and diffusionLesson25 min
- Deep learning architectures quizQuiz15 min
Large language models
- What a large language model is, next-token predictionLesson30 min
- Tokens and tokenisationLesson30 min
- Pretraining on text at scaleLesson30 min
- From base model to assistantLesson30 min
- The context window, and how a chat really worksLesson30 min
- What LLMs are bad at, and the model landscapeLesson35 min
- Large language models quizQuiz15 min
Building with LLMs
- Prompt engineering that worksLesson30 min
- Structured output and how to validate itLesson30 min
- Retrieval-augmented generation end to endLesson35 min
- Tool use and agentsLesson30 min
- Evaluating LLM applicationsLesson35 min
- Cost, latency, model choice and a worked designLesson35 min
- Chapter quiz: Building with LLMsQuiz15 min
- Lab: build the retrieval half of RAG from scratchLab40 min
AI security and safety
- The attack surface of an AI systemLesson25 min
- Prompt injection, direct and indirectLesson30 min
- Layered defences for LLM applicationsLesson30 min
- Poisoning, supply chain risk and adversarial examplesLesson30 min
- Privacy risks and model theftLesson30 min
- OWASP Top 10 for LLMs, red teaming and threat modellingLesson35 min
- Chapter quiz: AI security and safetyQuiz15 min
Responsible AI and production
- From notebook to productLesson30 min
- Monitoring, drift and when to retrainLesson30 min
- Fairness and how to measure itLesson35 min
- Transparency, explainability and human oversightLesson30 min
- Regulation and the wider costs of AILesson35 min
- Your next steps in AILesson30 min
- Chapter quiz: Responsible AI and productionQuiz15 min