The AI Blueprint – Understand AI. Evaluate AI. Apply AI.

Objective:

Understand how machines learn from data, conceptually and practically, so you can work confidently alongside ML-powered systems without needing to build them yourself.

Topics: 

• What machine learning is: finding patterns in data to make predictions about new data
• Supervised learning: teaching with labelled examples — spam filters, fraud detection, credit scoring
• Unsupervised learning: finding hidden patterns without labels — customer segments, anomaly detection
• Reinforcement learning: learning by trial and reward — recommendation engines, game-playing AI
• Training, validation, and testing: why models must be evaluated before they are trusted
• Overfitting and bias: how models go wrong and why human oversight always matters
• Neural networks explained simply: layers of pattern recognition with no mathematics required

Objective:

Understand how Generative AI works at a conceptual level so you can use it confidently, evaluate its outputs critically, and explain it clearly to others.

Topics: 

• What makes AI generative: it creates new content, it does not just retrieve existing content
• Large Language Models explained: trained on vast text, predicting the most useful next word
• Tokens, context windows, and temperature — what these mean without any technical background
• What Gen AI produces: text, summaries, plans, analysis, code, images, audio, and video
• Multimodal AI: working across text, images, documents, and voice simultaneously
• The main models today: ChatGPT, Claude, Gemini, Llama — key differences in plain language
• Hallucinations: why AI produces confident but incorrect outputs and how to catch them

Objective:

Master the prompting principles that separate specific, useful AI outputs from generic, unusable ones — the single skill that makes every other AI capability more powerful.

Topics: 

• Why prompting is a learnable skill and not a matter of luck or guessing
• The 4-part prompt structure: Role, Context, Task, Format — with plain everyday examples
• Zero-shot prompting: asking AI to do something with no examples provided
• Few-shot prompting: showing AI two or three examples before asking it to produce its own
• Chain-of-thought prompting: asking AI to reason step by step for more reliable answers
• Prompt chaining: breaking a complex task into a sequence of simpler AI steps
• Common failure patterns: why prompts produce bad outputs and exactly how to fix each one
• Building a personal prompt library: starting with what you learn today, growing every week

Objective:

Understand the shift from using AI as a tool you prompt to deploying AI agents that work continuously on your behalf — and what this means for every professional role.

Topics: 

• The critical difference: Gen AI does what you ask, Agentic AI does what you need without being asked
• What an AI agent is: goals, memory, available tools, decision loops, and degrees of autonomy
• How agents plan: breaking a goal into steps and executing them independently
• Tool use: how agents connect to email, calendars, databases, documents, and other systems
• Multi-agent systems: multiple specialised agents collaborating on a complex shared task
• Human-in-the-loop: where well-designed agents stop and wait for human review or approval
• What agents can and cannot do reliably today — honest expectations, no hype
• Where Agentic AI is heading over the next 12 to 24 months

Objective:

Understand how AI models go from a promising experiment to a reliable system running in production at scale — and why every professional working near AI needs to understand this pipeline.

Topics: 

• The gap between building an AI model and running it reliably at scale in the real world
• MLOps: versioning models, automating pipelines, monitoring performance, and detecting drift
• Model drift: why AI performance degrades over time as the world changes, and what is done about it
• CI/CD for AI: how updates to models are tested and deployed safely without breaking production
• LLMOps: how managing large language models differs from traditional ML operations
• Fine-tuning vs RAG vs prompt engineering: three ways to customise AI for your specific context
• Evaluation frameworks: how AI outputs are measured for quality, accuracy, safety, and fairness
• Governance and compliance: auditability, data privacy, and responsible AI in production systems

Objective:

Connect every concept from the course into a single coherent mental model so you can navigate the AI landscape confidently, understand where each technology fits, and know what to learn next.

Topics: 

• The AI stack connected: data → ML models → Gen AI → Agents → MLOps/LLMOps keeping it running
• How each layer enables the next: why you cannot have reliable agents without solid foundations
• Reading the landscape: where major AI products and platforms sit in this framework
• The professional’s map: where your specific role intersects with each layer
• What to learn next: how AI Foundations connects directly into your role-specific course
• Questions to ask when someone presents you with an AI initiative, product, or proposal