The Technology Learning Journey · Level 7 of 8

Artificial Intelligence and Intelligent Automation

Modules 81–90 · 3 courses · 50 lessons · 6–8 weeks

Study this level

All modules ready to study

Level 7 tests

Pass every module quiz above to unlock these. Passing them unlocks the next level.

Level 7 testArtificial Intelligence · questions from this level · 20 questions · 100 marksNot taken
Levels 1–7 combined testQuestions from Levels 1 to 7 · 20 questions · 100 marksNot taken

Level 7 project: Knowledge-Based AI Assistant. The tests check your understanding. The project shows you can build it, debug it and explain your design choices, so do it alongside the tests, not instead of them.

The full syllabus

Not just prompting: what models do, what they can't reliably do, and how to control and evaluate them. User → backend (permissions, context) → model → knowledge retrieval → controlled tools with human approval → checked answer or authorised action.

Course 7.1 Foundations

81 Introduction to AI: history; machine learning; deep learning; generative AI; uses and limits.
82 Large Language Models: tokens; training vs inference; transformers; context windows; uncertainty.
83 Prompt Engineering: instructions and context; few-shot; structured output; templates; testing.

Course 7.2 AI Application Engineering

84 AI APIs: API concepts; keys; streaming; cost and speed; error handling.
85 Embeddings and Vector Search: embeddings; similarity; vector indexes; semantic search; quality.
86 RAG: architecture; chunking; retrieval and reranking; grounding and citations; evaluation.
87 Agents and Tool Calling: function calling; workflows; tool permissions; MCP; human approval.

Course 7.3 Automation, Evaluation and Governance

88 Workflow Automation: triggers; orchestration; n8n; scheduled and event workflows; approvals.
89 Evaluation and Safety: hallucination; prompt injection; test sets; privacy; monitoring.
90 AI Application Project: risk; knowledge prep; integration; safeguards; deployment.

Final project: Knowledge-Based AI Assistant (12–20 hours). Answers from approved material, says when it doesn't know, cites sources, uses only permitted tools. Checklist: define who it serves and what it must never do; prepare and chunk the approved knowledge; build retrieval and check it finds the right passages; connect a model API with keys kept secret; show sources with every answer; say "I don't know" when the material doesn't cover it; give tools the smallest permissions and require approval for any action; test against a planted prompt-injection document; build a small test set and measure correctness; log questions and answers without storing unnecessary personal data; deploy behind login; document limits and risks.

Assessment: fundamentals 15%, API integration 15%, retrieval and grounding 20%, evaluation and safety 25%, final app 25%. Pass 70% with grounded answers, honest uncertainty and restricted tools.