This repository collects teaching material, notebooks, and references used in the Master-level Machine Learning course led by SITRAKA.

• 00_Introduction/ — onboarding notebooks, Python/Colab setup, and early LLM references.
• 01_Fundamentals/ — core Machine Learning concepts, algorithm summaries, and flashcards.
• 03_Advanced/ — deep learning examples, TensorFlow modelling, and orchestration with Airflow.
• docs/ — supplementary theory, formulas, and detailed notes.
• resources/ — datasets, external references, and enrichment materials.
• LOGISTIC_REGRESSION_DDEFI.ipynb — practical logistic regression notebook.

Students completing this course should be able to:

• Explain the main principles of supervised, unsupervised, and deep learning.
• Design and evaluate end-to-end ML workflows for business use cases.
• Interpret model behaviour through metrics, regularization, and bias-variance tradeoffs.
• Connect modern NLP architectures such as transformers to practical applications.
• Understand how orchestration tools like Airflow support production ML pipelines.

• 13/11 — LIRONE #1 (6h)
• 25/11 — SITRAKA #2 (6h) — Introduction to Machine Learning
• 28/11 — SITRAKA #3 (6h) — ML, Deep Learning, and collaborative work

• 19/12 — LIRONE #4 (6h)

• 29/01 — LIRONE #5 (6h)
• Additional sessions: TBD

• Data preparation: cleaning, missing values, duplicates, outliers
• Feature engineering: scaling, encoding, feature creation
• Feature selection: signal versus noise
• Bias-variance tradeoff and model complexity
• Evaluation metrics and selection of the right metric

• Linear regression
• Logistic regression
• Tree-based models: decision trees, random forests, gradient boosting

Business use cases: marketing mix modeling, client scoring, churn prediction, sports analytics

• Clustering: K-Means, hierarchical clustering
• Dimensionality reduction: PCA

Business use cases: customer segmentation, targeting strategies

• Differences between traditional ML and deep learning
• Neural network training and backpropagation
• Loss functions such as cross-entropy

Business use cases: computer vision for tracking, advanced predictive models

• NLP fundamentals and model evolution
• Word embeddings and attention mechanisms
• Transformer architectures in modern applications

Business use cases: finance analytics, automated report generation, sentiment analysis

• Feature Engineering Repository
• Designing Machine Learning Systems

• Synthetic data generation
• Agent-Based Modelling (ABM)
• Quantitative finance trends and research

• NotebookLM
• Diagrams (architecture visualization)

• Add Finance NLP use cases
• Add datasets and data exploration pipelines
• Add an end-to-end ML project
• Add MLOps and deployment examples

Machine Learning = Data + Models + Business Impact