What to study for each CAIP domain
Fundamental concepts and principles of artificial intelligence
Covers the fundamental concepts and principles of artificial intelligence, including what AI is, narrow and general AI, and how machine learning and deep learning relate to AI. It introduces expert systems, common applications such as recommendations and voice assistants, and the basic model life cycle from problem definition through training to inference.
Key skills
AI definitions and scope, narrow AI, machine learning concepts, deep learning concepts, expert systems, AI applications, problem definition, training and inference
Data analysis and Visualization
Covers data analysis and visualization in support of AI projects, including collecting and understanding data, handling missing values and outliers, and preprocessing such as encoding and scaling. It addresses exploratory data analysis and choosing appropriate charts, including histograms, bar charts, scatter plots, and correlation views, to communicate patterns clearly.
Key skills
data collection and quality, missing value handling, outlier treatment, preprocessing and encoding, feature scaling, exploratory data analysis, histograms bar and scatter charts, correlation analysis
Building Machine Learning models
Covers building machine learning models, including supervised classification and regression, unsupervised clustering, and reinforcement learning. It addresses training, validation, and test data, overfitting and how to control it with regularization and cross-validation, evaluation metrics for imbalanced data, and common models such as decision trees and random forests.
Key skills
classification and regression, clustering with K-means, reinforcement learning, train validation and test splits, overfitting, regularization and cross-validation, evaluation metrics, decision trees and random forests
Deep Learning and Natural Language Processing
Covers deep learning and natural language processing, including neurons and perceptrons, activation functions, backpropagation, and convolutional and recurrent architectures. It also covers transformer models and self-attention, and core NLP work such as tokenization, word embeddings, machine translation, sentiment analysis, and speech recognition.
Key skills
perceptrons and activation functions, backpropagation and gradient descent, CNNs and RNNs, transformers and self-attention, tokenization and embeddings, machine translation, sentiment analysis, speech recognition
Knowledge and application of Computer Vision and Robotics
Covers knowledge and application of computer vision and robotics, including image classification, object detection, semantic segmentation, object tracking, pose estimation, and OCR. It also addresses robot perception, sensors and actuators, vision-guided grasping and navigation, and depth estimation for autonomous systems.
Key skills
object detection and segmentation, object tracking, pose estimation, OCR and facial recognition, robot perception, sensors and actuators, navigation and grasping, depth estimation
AI Security
Covers AI security, including adversarial examples, data poisoning, model extraction, and prompt injection, together with defences such as adversarial training and input validation. It also addresses privacy leakage from models, protecting model artefacts, authenticated and rate-limited model access, and monitoring deployed models for drift and abuse.
Key skills
adversarial examples, data poisoning, model extraction, prompt injection, adversarial training, input validation, privacy leakage, model artefact integrity, API access control, drift monitoring
AI Ethics, Governance, Strategy
Covers AI ethics, governance, and strategy, including algorithmic bias and fairness, explainability, transparency, and human oversight of high-impact decisions. It addresses governance accountability, model documentation, data governance, third-party AI risk, aligning AI initiatives with organizational strategy, and acting when monitoring shows unfair outcomes.
Key skills
bias and fairness, explainability and transparency, human oversight, governance accountability, model cards and documentation, data governance, third-party AI risk, AI strategy alignment