Master the entire scope of KOAI Subject 1 (Foundational Skills and Classical ML) through theory and hands-on code. Using NumPy·Pandas·scikit-learn, you build regression, classification, and clustering pipelines end to end and cover model evaluation and feature engineering. It is the essential course that serves as the starting point for every KOAI entrant.
Published: May 16, 2026 | Last updated: May 16, 2026 · Based on the KOAI 2026 guidelines
Track
Foundation
Required for all KOAI applicants
Target Grade
Grade 9 to Grade 11
Prior Python study required
Recommended Hours
About 10 hours
For 1:1 · varies 6–14 hours
KOAI Mapping
Full scope of Subject 1
Syllabus 1-1 to 1-4
After completing F1, the student Build regression, classification, and clustering pipelines end to end, from scratch, with NumPy, Pandas, and scikit-learn. This isn't just about learning how to call libraries — the goal is to understand theoretically why a model behaves the way it does, interpret model evaluation metrics, and reach the level where you can improve performance through feature engineering.
F1 does not teach Python itself. It assumes you've already learned basic Python syntax and dives straight into the core of data science and machine learning from the very first session. This course covers the entire scope of KOAI Subject 1 (syllabus 1-1 through 1-4) without gaps, making it the common starting point for every KOAI candidateIt is.
Prior learning of basic Python syntax is Requiredrequired. F1 doesn't teach Python and goes straight into NumPy/Pandas/scikit-learn. If your Python is still shaky, the CIT Coding Class Python Basicsfirst, or build the foundation with the elementary track's E2·E3 first.
Below is the standard plan based on 1:1 instruction. Depending on the student's prior knowledge and pace of absorption, some weeks are accelerated, compressed, or covered in greater depth. Core tools: NumPy, Pandas, Matplotlib/Seaborn, scikit-learn, XGBoost/LightGBM, Optuna, Kaggle.
| Week | Topic | Key Deliverable |
|---|---|---|
| 1 | NumPy vectors & matrices, Pandas DataFrame | EDA notebook (Kaggle Titanic) |
| 2 | Matplotlib/Seaborn, data preprocessing | EDA notebook (Kaggle House Prices) |
| 3 | Linear regression (theory & implementation) | Regression baseline + residual analysis |
| 4 | Logistic regression, L1/L2 regularization | Classification baseline |
| 5 | K-NN, decision trees | Model comparison notebook |
| 6 | Ensembles (Random Forest, Gradient Boosting) | Applying XGBoost/LightGBM |
| 7 | SVM | SVM kernel comparison experiment |
| 8 | K-means, PCA | Dimensionality reduction + cluster visualization |
| 9 | t-SNE, UMAP, DBSCAN | Embedding visualization notebook |
| 10 | Model evaluation metrics, confusion matrix, ROC | Metrics interpretation report |
| 11 | Underfitting & overfitting, cross-validation | Learning curve analysis |
| 12 | Hyperparameter tuning (Grid, Random, Optuna) | Tuning experiment notebook |
| 13 | Feature engineering (sliding window, encoding, statistical moments) | Performance improvement comparison from added features |
| 14 | Mini Kaggle competition (capstone) | Leaderboard submission + retrospective |
※ Weeks are content units; actual time varies by student. Recommended about 10 hours, ranging 6–14 hours.
Every week you write a Jupyter notebook. English notebooks include parallel Korean annotations for key terms. This leverages international-school students' strength in English while simultaneously preparing for the Korean-language essay portion of the KOAI Round 2 exam.
You submit directly to one public Kaggle competition and write a 5-page analysis and retrospective report. Your leaderboard score and the improvement process become portfolio assets as they are.
The portfolio — at 40%, the largest weight in the KOAI Round 1 application review — cannot be built in the few weeks right before the exam. Every deliverable from F1 is designed to stack up as a cumulative asset.
GitHub
1 organized repo koai-classical-ml
Kaggle
Active profile + competition submission history
Notion
4–6 learning journal entries (material for the personal statement)
This track record accumulates as dated evidence in the KOAI Round 1 application's Portfolio 40% · AI competency 30% item. The earlier you start, the more depth you'll have by exam time.
F1 is the first course in the foundation track of the KOAI curriculum. For the full track structure, see KOAI Prep Curriculum Hub.
Current Course
F1. Foundations I
Python + Classical ML
Yes. F1 assumes you've already learned basic Python syntax (variables, loops, functions, lists, dictionaries), because it doesn't teach Python itself and goes straight into NumPy, Pandas, and scikit-learn. If your Python is still shaky, take the CIT Coding Class Python Basics first, or build the foundation with E2/E3 in the elementary track.
F1 maps to the entire scope of KOAI syllabus Subject 1 (Foundational Skills and Classical ML), from 1-1 through 1-4. It covers everything from data preprocessing and EDA to linear/logistic regression, trees and ensembles, SVM, dimensionality reduction and clustering, model evaluation, hyperparameter tuning, and feature engineering.
No. The roughly 10 hours is a recommendation based on 1:1 instruction and varies by ±30–50% depending on the student's prior knowledge and pace of understanding (roughly 6–14 hours). Students with prior ML experience may need less than the recommended time, while beginners may need more. We calculate an individual time plan for each student in the first diagnostic session.
After F1 comes F2 (Foundation II: Neural Networks & Deep Learning). High-school division candidates then go deeper with A1 (Computer Vision) and A2 (NLP & Audio), while middle-school division candidates cover F1 and F2 in a condensed form in the M1 Middle School Comprehensive Class. The portfolio is prepared in C1, exam execution in C2, and interviews in C3.
F1 leaves behind a clean GitHub repo (koai-classical-ml), an active Kaggle profile, and 4–6 Notion learning journal entries as cumulative assets. This track record serves as date-stamped evidence for the Portfolio 40% and AI Competency 30% items in the KOAI Round 1 application. Since the portfolio is a function of time, the earlier you start, the deeper it gets.
We'll individually design F1's start timing and time plan in a diagnostic session, tailored to your child's prior knowledge and goals.