16 lessons · Chromebook & mobile ready
Advanced AI
Build, evaluate, and audit AI systems — beyond prompting
Move from prompting to building — train, evaluate, retrieve, audit, and ship AI systems like a creator, not just a consumer.
For learners & program leaders
What you'll walk away with
Grades 10–12 (strong 9ths welcome) who already understand AI basics and want a specialty pathway beyond chat tools. Flexible schedule — 16 sessions over 8 weeks.
Note: Recommended: AI Literacy plus Python & AI Foundations (or equivalent coding basics).
- Frame AI problems with clear tasks, metrics, and human review
- Work with training data — features, labels, leakage, and dataset bias
- Train classifiers and read confusion matrices, precision, and recall
- Explain LLMs beyond prompting: tokens, embeddings, RAG, and agents
- Choose among prompt, RAG, and fine-tune — then evaluate with a harness
- Audit fairness and AI security risks, then ship and defend a capstone system
Built as a specialty pathway
Designed above AI Literacy and beside Python & AI — aligned to CSTA AI specialty / creator outcomes (data, ML, evaluation, and impact), not a watered-down survey of chatbots.
Skills · Capstone · Standards
ML evaluation labs · RAG & agents · Fairness audits · AI security awareness · Ship & defend a system
Capstone: Demo, Audit & Defend · Built to align with CSTA K–12 CS Standards · standards packet available
Inside the path
16 lessons · 8 weeks
Choose a week to see what each session teaches. Pace flexes to your classroom, family, or program.
01 · Learn it
Coach’s note, quick explainer, and word help — the goal and big idea in plain language.
02 · Do the activity
Guided practice with instant feedback, then a from-scratch challenge and check-for-understanding.
03 · Reflect
Ethics or reflection moments on AI, data, cyber defense, or digital citizenship.
04 · Earn XP & badges
Progress shows on a clear roadmap — for learners, families, and instructors.
Week 1 of 8
Frame & Data
Problem framing, task types, features/labels, leakage, and dataset bias.
01
Session 1
AI Systems: Framing Problems Worth Solving
Define a problem worth solving with AI — task type, success metric, and where a human stays in the loop.
02
Session 2
Data for Machine Learning
Prepare data for learning: features, labels, leakage risks, and how biased datasets shape unfair models.
Get started
Try a lesson — then talk to us
Learners and families can demo the real lesson canvas first. Schools and programs: tell us your timeline and we'll help you pilot a class, after-school block, or Scout troop.
Families & learners
Demo with no account. Then reach out for a self-paced class code or live support.
Schools & admins
Class codes, roster progress, and standards docs — request a pilot when you're ready.
