Module 1: Python & Data Foundations for ML (Weeks 1–4)

Objective: Ensure comfort with Python, data preprocessing, and statistics — critical for every ML model downstream.

Weeks 1: Python for Machine Learning

● Jupyter Notebooks & Python basics

● Control structures, functions, imports

● NumPy arrays and broadcasting

● Pandas DataFrames: manipulation and cleaning

Weeks 2: Data Wrangling & Feature Engineering

● Handling missing data

● Outlier detection & imputation strategies

● Encoding categorical variables (one-hot, label)

● Scaling & transformation (standardization, normalization, log transform)

Weeks 3: Exploratory Data Profiling

● Summary statistics

● Correlation and feature relationships

● Skewness & distributions

● Visual checks using Matplotlib/Seaborn

Weeks 4: Train/Test Splits & Workflow Basics

● Cross-validation fundamentals 

● Data leakage pitfalls

● Pipeline overview in scikit-learn

● Evaluation metrics: accuracy, MAE, RMSE, R², F1

Project: Clean, analyze, and prepare a dataset for modeling (e.g., Titanic, House Prices)

Module 2: Core Machine Learning – Supervised & Unsupervised (Weeks 5–16)

Objective: Master supervised and unsupervised learning with solid understanding of model choices, evaluation, tuning, and practical usage.

Weeks 5–7: Regression Techniques

● Linear Regression & assumptions

● Polynomial Regression

● Regularization: Ridge & Lasso

● Model interpretation: coefficients, residuals, bias-variance

Weeks 8–10: Classification Algorithms

● Logistic Regression

● K-Nearest Neighbors (KNN)

● Decision Trees & entropy/gini

● Evaluation: ROC-AUC, confusion matrix, precision-recall curves

Weeks 11–13: Ensemble Learning

● Bagging & Random Forest

● Boosting: Gradient Boosting, AdaBoost

● Hyperparameter tuning via GridSearchCV

● Feature importance & SHAP (intro)

Weeks 14 - 16: Unsupervised Learning

● K-Means clustering

● Hierarchical clustering

● PCA for dimensionality reduction

● Silhouette score, elbow method

Project: Compare models on a business problem (e.g., customer churn or loan default)

Module 3: Advanced ML Concepts & Project Delivery (Weeks 17–24)

Objective: Focus on building robust ML pipelines, tuning models, interpreting outputs, and delivering real-world ML projects.

Weeks 17–18: Model Selection & Evaluation Strategies

● Cross-validation techniques

● Bias-variance tradeoff deep dive

● Overfitting/underfitting analysis

● Stratified sampling & class imbalance handling (SMOTE, weighting)

Weeks 19–20: Pipelines & Production Readiness

● Full pipeline creation in scikit-learn

● Custom transformers with FunctionTransformer

● Model persistence (Pickle/Joblib)

● Version control of models & data

Weeks 21–22: Explainable ML (XAI)

● Feature importance methods

● Partial dependence plots (PDPs)

● SHAP & LIME for model explainability

● Trust and fairness in ML

Weeks 23–24: Capstone Hackathon

● 2-week real-world ML challenge

● Full cycle: data cleaning -> feature engineering -> modeling -> evaluation -> reporting

● Peer code review and live presentations

Capstone Examples:

● Predictive maintenance for machinery

● Price prediction for online retail

● Credit risk scoring 

Tools & Environment

● Python Libraries: Pandas, NumPy, Scikit-learn, Matplotlib, Seaborn

● Dev Tools: Jupyter Notebook, Git/GitHub

● Google Colab, Visual Studio Code, AWS

Objective: To continue from the AML foundation, focusing on advanced techniques, end-to-end production readiness, model deployment, GenAI, MLOps, and Big Data applications.

Module 4: Deep Learning Essentials (Weeks 25–32)

• Neural networks, backpropagation, activation functions

• CNNs (VGG, ResNet) for image classification

• RNNs, LSTMs, GRUs for sequential/time series data

• Autoencoders, Variational Autoencoders (VAEs)

• GANs – Generative Adversarial Networks

Capstone: Object detection OR stock prediction

Tools: TensorFlow, Keras, PyTorch, Google Colab

Module 5: Advanced NLP & Transformers (Weeks 33–36)

• Word embeddings (Word2Vec, GloVe, FastText)

• Attention Mechanism, Transformers intro

• ERT, RoBERTa, DistilBERT fine-tuning

• Hugging Face APIs and LangChain for LLMs

Capstone: LLM-powered chatbot or QA system

Tools: Hugging Face, LangChain, OpenAI APIs, spaCy

Module 6: Generative AI & Large Language Models (Weeks 37–40)

• Introduction to Generative AI and foundational concepts

• Hands-on with OpenAI GPT models: text generation, summarization, code generation

• Building RAG (Retrieval Augmented Generation) pipelines

• LangChain for custom LLM workflows (tools, memory, agents)

• Document intelligence and knowledge base augmentation

Capstone: RAG-based GenAI application (e.g., legal Q&A bot, contract summarizer)

Tools: OpenAI API, Hugging Face Transformers, LangChain, LlamaIndex, Pinecone, Weaviate

Module 7: MLOps & Cloud Deployment (Weeks 41–44)

• CI/CD for ML – GitHub Actions, ML lifecycle

• Model versioning and tracking using MLflow

• Docker, FastAPI, and Streamlit for model serving

• Monitoring (Prometheus, Grafana)

Capstone: Deploying full ML/GenAI pipeline to AWS/GCP

Tools: MLflow, Docker, FastAPI, Streamlit, AWS/GCP

Module 8: Big Data & Real-Time Pipelines (Weeks 45–46)

• PySpark for distributed ML

• Data pipeline design with Airflow

• Real-time ingestion with Kafka

• Scalable storage: AWS S3, Databricks

Capstone: Streaming-based ML/GenAI integration

Tools: Apache Spark, Airflow, Kafka, Databricks

Module 9: Responsible AI & Final Capstone (Weeks 47–48)

• Model interpretability: SHAP, LIME, PDPs

• Fairness & bias mitigation techniques in traditional & GenAI models

Final Capstone: End-to-end real-world AI solution

• Live presentation before industry panel

Tools: SHAP, LIME, AIF360, Google Cloud, LangChain

Modules list will be displayed here.

Student reviews will be shown here.