Artificial Intelligence & Machine Learning
Konsep AI, Machine Learning, Deep Learning, dan implementasinya
Pengantar AI & Machine Learning
GratisMateri pembuka: definisi AI/ML/DL, 4 tipe pembelajaran (supervised/unsupervised/reinforcement/self-supervised), aplikasi nyata, dan peta 15 materi kurikulum.
Sejarah & Evolusi AI
GratisDari Turing Test 1950 hingga ChatGPT 2022. Timeline lengkap, pola Spring/Winter, dan 6 gelombang AI saat ini termasuk Foundation Models, Agentic AI, dan Open Weights.
Matematika & Statistik untuk AI
Terkunci4 pilar matematika AI: Linear Algebra (vektor & matriks), Calculus (gradient descent), Probability, Statistics. Plus information theory & strategi belajar pragmatis.
Supervised Learning — Regresi & Klasifikasi
TerkunciTipe ML paling populer. Linear & Logistic Regression, workflow 7-tahap, common pitfalls (data leakage, overfitting, class imbalance), dan studi kasus credit scoring.
Unsupervised Learning — Clustering & Dimensionality
TerkunciK-Means, Hierarchical, DBSCAN. PCA & t-SNE untuk reduksi dimensi. Anomaly detection. Plus studi kasus customer segmentation Tokopedia 100 juta user.
Algoritma ML Klasik — Decision Tree, RF, SVM, KNN
TerkunciTools production yang masih dominan: Decision Tree, Random Forest, XGBoost/LightGBM (king tabular), SVM, KNN, Naive Bayes. Cara memilih algoritma per situasi.
Neural Networks & Deep Learning Basics
TerkunciPerceptron, multi-layer network, activation functions (ReLU/GELU/Softmax), backpropagation, optimizers (Adam), dan regularization. Studi kasus AlphaFold.
Model Evaluation & Validation
TerkunciConfusion matrix, precision vs recall, F1, ROC-AUC. Metric regresi (MAE, RMSE, R²). Cross-validation, stratified K-Fold, overfitting detection. Studi kasus medical.
Feature Engineering & Data Preprocessing
Terkunci80% pekerjaan ML. Handling missing data, encoding (label/one-hot/target), scaling, feature creation, outlier handling. Studi kasus Kaggle dengan 800+ features.
CNN & Computer Vision
TerkunciConvolution, pooling, arsitektur klasik (LeNet→ViT). Tugas CV: classification, detection, segmentation. Transfer learning + studi kasus visual search JD.ID.
NLP, RNN, LSTM, Transformer
TerkunciWord embedding (Word2Vec), RNN, vanishing gradient, LSTM gates, attention mechanism, dan revolusi Transformer 2017. BERT, GPT, T5. Studi kasus Google Search.
Generative AI & Large Language Models
TerkunciLLM, prompt engineering (zero/few-shot, CoT), RAG (Retrieval Augmented Generation), fine-tuning, AI Agents, dan diffusion models. Build RAG application di tugas.
MLOps & Deployment
TerkunciPipeline end-to-end, tooling (DVC, MLflow, Airflow, FastAPI, Evidently), deployment patterns (batch/real-time/edge), monitoring drift, CI/CD/CT. Studi kasus Gojek.
AI Ethics, Bias & Responsible AI
Terkunci6 tipe bias, fairness metrics, mitigation strategies, transparency (SHAP, LIME), privacy (differential privacy, federated learning), dan regulasi global (GDPR, EU AI Act, UU PDP).
Future of AI — AGI, Multimodal, Capstone Project
TerkunciMateri penutup: 6 mega-trend (agentic, multimodal, on-device, open weights, embodied, AI for science), AGI, AI safety, karier, dan tugas akhir capstone project end-to-end.