AI literacy
Reproduce a key calculation with guided notebooks and synthetic data, then explain it in two minutes.
한국어 안내
16-week high-school studio
AI + ManagementFrom model to market.
A 16-week studio connecting model performance to user value, product KPIs, and responsible launch decisions. Available online or in person in Apgujeong, in 1:1 and small-group (1:n) formats.
Core questionWhat evidence supports the product decision, beyond the model score?
16 weeksone studio each week
72 hoursguided + independent
4 levelsreadiness-based routes
AI and management
are not two subjects.
Here, AI is not simply a tool that produces an answer. It is a system students evaluate before making a product decision.
Students frame the problem, inspect data, build a baseline, and evaluate a model. They connect the evidence to a PRD, KPIs, cost, risk, and a launch decision.
50%AI, data, models
30%product, business
20%ethics, teamwork
- ProblemIdentify whose problem matters
- DataRecord source, unit, quality, limits
- Baseline + AICompare on the same evaluation set
- Product KPITest the user task and operating cost
- DecisionChoose go, iterate, or stop
Four readiness levels
Start by readiness,
not by grade.
Every student works through the same 16-week problems and safety requirements. Implementation depth, tools, and independence change.
Grades 9-12 are the recommended age range. Placement follows a diagnostic, and a student can change routes during the term.
Evidence data notebook, simple app, two-minute explanation
ML and MVP
Build a small model and MVP, then revise a PRD-lite and sprint from user-task evidence.
Model and product analytics
Compare subgroups, sensitivity, model KPIs, and product KPIs in a decision memo.
Product, venture, responsibility
Manage scope and risk gates, then defend a bounded launch decision for a capstone.
16-week curriculum
Learn, build,
then decide.
Each week includes a 180-minute guided studio and 90 minutes of independent work. Detailed lesson materials are available in the tutor and student LMS.
Problem, execution, data, baseline
- W01What AI should and should not solve
- W02Python, Jupyter, and team operations
- W03Data literacy and an EDA memo
- W04ML baseline and success metric
Evaluation, MVP, Korean NLP, decisions
- W05Confusion matrix and metric decision
- W06App Inventor MVP and user test
- W07Stanza Korean NLP and user value
- W08Search, games, and trade-off decisions
Performance, responsibility, safety
- W09Deep learning cost-performance lab
- W10Robustness, bias, and model cards
- W11mini-HELM and product KPIs
- W12Age-appropriate AI safety audit
Business judgment and capstone studio
- W13Opportunity, customer, value, cost
- W14Capstone sprint and baseline
- W15Validation and launch decision
- W16Demo day, pitch, postmortem
Beyond a list of activities.
Evidence for every decision.
Artifacts show what a student compared, why a choice was made, and what evidence would change the decision.
- Data and model evidenceEDA memo, non-AI baseline, ML and NLP experiments, confusion matrix, subgroup evaluation
- Product evidenceProblem statement, PRD-lite, MVP, user-task test, product KPIs
- Decision evidenceMetric decision, model card, safety audit, business and launch memos
- Operations evidenceGitHub issues and PRs, sprint board, reviews, AI-use log
- Capstone evidenceReproducible repository, evaluation report, five-minute demo, pitch, postmortem
Different interests.
The same evidence standard.
- Korean AI Study Assistant Evaluator
- Sports Team Decision Lab
- Local Business AI Product
- School Sustainability AI
- Accessible School App
The learning record parents can review
- Weekly artifacts and the next improvement question
- Student-written decision memos and reflections
- GitHub evidence connecting code, data, and evaluation
- What AI assisted and what the student decided
- Final demo, pitch, postmortem, and next roadmap
Safety boundaries applied to every project
- No collection of students' personal, sensitive, face, location, health, home, or grade data
- Public, synthetic, and aggregate data with source and version records
- No products that automatically score, profile, or monitor students
- Explicit uncertainty, refusal, human help, retention, and deletion paths
- An AI-use log plus live student explanation and modification
Standards and sources
Official frameworks
set the boundaries.
We do not copy university courses. Official sources define concepts and tool boundaries; CIT writes original high-school explanations, synthetic data, problems, and rubrics.
Last reviewed: 2026-08-13 / Official documents and tool status are checked before teaching
Education and competence
- AI4K12Grade 9-12 progression and the Five Big Ideas
- CSTA StandardsAI, data, ethics, and human-centred collaboration
- UNESCO AI Competency FrameworkStudent agency and responsible co-creation
Product and execution
- EntreCompIdeas and Opportunities, Resources, Into Action
- MIT App InventorSmall AI MVPs and user flows
Models and risk
- Stanford StanzaKorean NLP pipelines and error analysis
- Stanford HELMMulti-metric, transparent, reproducible evaluation
- NIST AI RMFRisk mapping, measurement, response, and accountability
Questions to answer
before starting
A student's current Python experience, interests, and independent work habits help us choose the right entry route.
01Can a student with no Python experience start?
Yes. Foundation begins with guided notebooks and fixed synthetic data. Students who can work with functions and tables can move into Applied, Advanced, or Studio work.
02Is placement fixed by grade?
No. The course serves Grades 9-12, but readiness matters more than grade. Students enter after a diagnostic and can change routes during the course.
03How are AI and management balanced?
The recommended balance is 50 percent AI and data, 30 percent product, project, and business decisions, and 20 percent ethics, teamwork, and communication.
04What can parents review?
Parents can review weekly data memos, notebooks, baseline comparisons, a PRD-lite, model cards, decision memos, and sprint records. The final package includes a repository, evaluation report, demo, pitch, and postmortem.
05Does the course use students' personal data?
No. Projects do not collect or profile students' personal, sensitive, face, location, health, home, or grade data. Public, synthetic, and aggregate data are preferred.
06Must students launch a real company?
No. The course does not promise revenue or launch. Students use evidence about user value, cost, risk, and product KPIs to choose a pilot, iteration, or stop decision.
Strong AI work starts
at the right level.
We review current experience and interests, then choose a route from Foundation through Studio.