Detailed Syllabus

Upper Secondary

Grades 9–12 90 minutes · 100 marks

Implementation and evaluation: the transformer pipeline, sampling and temperature, embeddings and bias, diffusion, and grounded agentic / RAG design - bridging to NOAI and IOAI / IAIO.

// who it's for

Students, teachers and coaches preparing for the most advanced IJAIO band - including learners in Upper Secondary and equivalent applied pathways.

// what success looks like

Learners explain end-to-end generative-AI systems, evaluate models with explicit metrics and trade-offs, and present defensible design choices for agentic or retrieval-grounded systems.

// where this band sits

The four-band journey

{{ c.band }}Grades {{ c.grades }}
{{ c.mode }}
{{ c.assess }}
// focus of this band

Implementation and evaluation - judging quality, fairness and design trade-offs against explicit criteria.

// learning outcomes

The five AI Pillars

Each Pillar leads with a disposition, then plain learning outcomes that combine what a learner knows and can do.

P{{ s.num }}

{{ s.title }}

Disposition. {{ s.disp }}
Learning outcomes
{{ o.n }}{{ o.t }}
Evidence. {{ s.evid }}
Key terms{{ kt }}
Assessment format
90 min
20 MCQ [40] + 3 structured / design / practical tasks [60]
Total 100 marks
Emerging-tech theme task each year · Python 3 + Jupyter / Colab
{{ a }}
// bridge to the higher olympiads

This band previews the ideas behind the National (NOAI) and International (IOAI / IAIO) Olympiads in Artificial Intelligence from first principles - by hand and in small Python - without the heavy mathematics or from-scratch engineering those olympiads add.

IJAIO → NOAI → IOAI
Hands-on Practice Lab
Play the Upper Secondary Practice Lab
Work with confusion matrices, thresholds and fairness, then deploy a classifier responsibly in one connected build.
🧪
// ready to practise
Try the Upper Secondary problem sets
Three interactive sets: advanced system design, grounding, metrics and guardrails - with clear marking.
// python notebook
Python Notebook · run the machinery in your browser
Eight runnable builds - embeddings and bias, a convolution, forward pass and loss, a trained classifier, temperature, RAG, and a model-card audit. Nothing to install.
>_
← Lower Secondary Overview