AI Technology Course

Lesson 1 • History and Evolution of AI

Traces AI from symbolic reasoning to modern deep learning milestones. Provides context for understanding why current techniques dominate the field.

Lesson 2 • Data as the Foundation of AI

Explains how data quality, quantity, and structure determine model performance. Establishes data literacy as a prerequisite for all subsequent chapters.

Lesson 3 • Types of Machine Learning

Distinguishes supervised, unsupervised, and reinforcement learning paradigms. Learners select the right paradigm for a given problem type.

Lesson 1 • Probability and Statistics

Grounds AI predictions in probabilistic reasoning and statistical inference. Enables interpretation of model confidence and uncertainty.

Lesson 2 • Calculus and Optimisation

Introduces derivatives, gradients, and optimisation landscapes used in model training. Gradient descent is explained as the engine of learning.

Lesson 3 • Linear Algebra Essentials

Covers vectors, matrices, and tensor operations central to neural network computation. Directly supports understanding of weight matrices and transformations.

Lesson 1 • Data Cleaning and Preprocessing

Addresses missing values, outliers, duplicates, and inconsistent formats. Clean data is the prerequisite for reliable model training.

Lesson 2 • Feature Selection Methods

Identifies the most predictive features using filter, wrapper, and embedded methods. Improves model interpretability and reduces overfitting risk.

Lesson 3 • Feature Engineering Techniques

Transforms raw variables into informative features through encoding, scaling, and creation. Strong features often matter more than algorithm choice.

Lesson 4 • Data Collection and Sourcing

Covers strategies for acquiring data from APIs, databases, and web sources. Establishes responsible collection practices aligned with privacy principles.

Lesson 1 • Clustering and Unsupervised Methods

Applies k-means, hierarchical clustering, and DBSCAN to discover hidden data structure. Evaluation without labels requires specialised metrics.

Lesson 2 • Regression Algorithms

Covers linear, polynomial, and regularised regression for continuous output prediction. Regularisation techniques prevent overfitting on real datasets.

Lesson 3 • Model Evaluation and Selection

Introduces cross-validation, confusion matrices, ROC curves, and model comparison frameworks. Rigorous evaluation prevents deployment of underperforming models.

Lesson 4 • Support Vector Machines

Explains margin maximisation, kernel tricks, and SVM classification and regression. Kernels extend SVMs to nonlinear decision boundaries.

Lesson 1 • Debugging and Improving Models

Diagnoses underfitting, overfitting, and training instability using systematic techniques. Practical debugging skills reduce wasted compute and accelerate iteration.

Lesson 2 • Neural Network Fundamentals

Explains neurons, activation functions, layers, and forward propagation. Provides the architectural vocabulary needed for all deep learning topics.

Lesson 3 • Training Deep Networks

Covers backpropagation, optimisers, batch normalisation, and dropout for stable training. These techniques are applied in every subsequent deep learning model.

Lesson 4 • Recurrent and Sequence Models

Covers RNNs, LSTMs, and GRUs for sequential and time-series data modelling. Sequence models precede transformer architectures introduced in the next chapter.

Lesson 1 • Generative Text Applications

Builds summarisation, translation, question-answering, and chatbot systems using LLMs. Connects transformer theory to production-ready NLP applications.

Lesson 2 • Text Preprocessing and Representation

Covers tokenisation, stemming, embeddings, and TF-IDF for converting text to numeric form. Representation quality directly determines downstream model performance.

Lesson 3 • Large Language Models and Fine-Tuning

Covers pretraining objectives, prompt engineering, and parameter-efficient fine-tuning methods. Students adapt foundation models to domain-specific tasks efficiently.

Lesson 4 • Classical NLP Tasks

Applies models to sentiment analysis, named entity recognition, and text classification. These tasks establish NLP baselines before transformer methods are introduced.

Lesson 1 • Multimodal AI Systems

Integrates vision and language through CLIP, image captioning, and visual question answering. Multimodal models represent the frontier of applied AI research.

Lesson 2 • Image Segmentation Techniques

Distinguishes semantic, instance, and panoptic segmentation with corresponding architectures. Segmentation enables pixel-level scene understanding for advanced applications.

Lesson 3 • Generative Vision Models

Covers GANs, VAEs, and diffusion models for image synthesis and editing. Generative models power creative AI tools and data augmentation pipelines.

Lesson 4 • Image Classification and Recognition

Applies pretrained CNNs and vision transformers to classification tasks with fine-tuning. Builds on CNN foundations from the deep learning chapter.

Lesson 1 • MLOps Pipelines and Automation

Builds automated training, validation, and deployment pipelines using MLOps principles. Automation reduces human error and accelerates the model update cycle.

Lesson 2 • Scalability and Infrastructure

Addresses horizontal scaling, GPU cluster management, and cloud-native AI infrastructure. Scalable infrastructure supports growing data volumes and user demand.

Lesson 3 • Responsible Production AI

Integrates fairness checks, explainability tools, and audit logging into deployment pipelines. Responsible practices reduce risk and build stakeholder trust.

Lesson 4 • Model Monitoring and Drift Detection

Detects data drift, concept drift, and performance degradation in live systems. Monitoring ensures models remain accurate after deployment conditions change.