Course Syllabus

Module 1: Introduction to AI & Python (3 hours)

Description: This module provides a foundation in AI/ML concepts and Python programming. It explains what AI and machine learning are (for example, that ML is a branch of AI focused on algorithms that let computers make decisions based on data ). Students learn to set up a Python development environment (Anaconda, Jupyter notebooks) and review basic Python syntax and data types. Emphasis is on writing clean Python code for data analysis. Key libraries are introduced, particularly NumPy for numerical data. By the end, learners understand how to load data and write simple scripts, preparing them for data manipulation in later modules.

Module 2: Data Handling & Visualization with Python (4 hours)

Description: This module teaches how to work with datasets in Python. Topics include data loading, cleaning, and exploratory data analysis using Pandas and NumPy. Students learn to manipulate dataframes (filter, merge, group-by) and handle missing values. The module also covers data visualization: creating plots (histograms, scatter plots, box plots) using Matplotlib and Seaborn to reveal patterns or anomalies. Instructors emphasize best practices for EDA. For example, as one course description notes, participants will “learn Python and Pandas for data manipulation” and “explore effective data visualization with Matplotlib and Seaborn”. Lab exercises involve plotting sample datasets and summarizing insights from charts.

Module 3: Supervised Learning (6 hours)

Description: Focuses on predictive modeling with labeled data. Begins with regression (predicting continuous outcomes) using linear regression, teaching how to fit a line to data and measure error (MSE). Then moves to classification (categorizing data) with algorithms like logistic regression, decision trees, and k-nearest neighbors. Students implement models in Python using scikit-learn, which is well-suited for these traditional ML tasks  . Each algorithm’s theoretical basis (e.g. how logistic regression outputs probabilities, or how a decision tree splits data) is explained. Lab sessions have students apply scikit-learn classifiers on example datasets (e.g. the Iris or Titanic dataset) to make predictions. Topics of training include handling categorical features, overfitting vs. underfitting, and model interpretation.

Module 4: Unsupervised Learning (3 hours)

Description: Introduces algorithms for unlabeled data. The module covers clustering techniques (e.g. K-Means, Hierarchical Clustering) where the goal is to group similar data points. It also introduces dimensionality reduction methods like Principal Component Analysis (PCA) to compress high-dimensional data. Using scikit-learn’s libraries, students learn to fit K-Means clusters to data and use PCA to visualize complex data in 2D. Theoretical concepts include distance metrics (for clustering) and variance capture (for PCA). Exercises involve clustering a sample dataset and plotting the clusters, illustrating how unsupervised methods find structure in data without predefined labels.

Module 5: Model Evaluation & Tuning (4 hours)

Description: Covers how to measure and improve model performance. Key topics include train/test split, cross-validation, and evaluation metrics (e.g. accuracy, precision/recall for classification; RMSE for regression). Students learn to compute confusion matrices and ROC curves for classifiers. The module also teaches hyperparameter tuning: how to use grid search or randomized search (e.g. scikit-learn’s GridSearchCV) to find optimal model settings. Additionally, building pipelines (using Pipeline in scikit-learn) is introduced to automate preprocessing and model training in one workflow. Labs involve tuning a decision tree or random forest by adjusting depth or number of trees and comparing performance. By the end, learners can reliably select and fine-tune models for best results.

Module 6: Introduction to Deep Learning (4 hours)

Description: This module transitions to neural networks and deep learning. Students learn the concept of a perceptron and multilayer networks trained by backpropagation. The module uses high-level libraries like TensorFlow (with Keras) and PyTorch to build simple neural networks. For example, a basic feedforward network for MNIST digit classification is implemented. TensorFlow is introduced as “an end-to-end platform for machine learning” that makes it easy to build and train models. Learners understand activation functions (ReLU, softmax), loss functions, and the training loop (gradient descent). The session compares frameworks: as one tutorial notes, TensorFlow and PyTorch “excel in deep learning and large-scale AI applications” (unlike scikit-learn, which is tailored to smaller-scale ML tasks)  . Through hands-on labs, students see how adjusting network layers and learning rates affects accuracy, gaining intuition for neural net training.

Module 7: Computer Vision Fundamentals (3 hours)

Description: Focuses on AI for image data. The module covers Convolutional Neural Networks (CNNs) – the core models for computer vision. It explains convolution and pooling layers, and how CNNs can detect patterns in images. Using TensorFlow or PyTorch, students build a simple CNN to classify images (for example, distinguishing cats vs. dogs or recognizing handwritten digits). Key concepts include image preprocessing (rescaling, normalization) and data augmentation. The module highlights practical tasks (image recognition/classification), which are classic AI problems – for instance, AI can automate the “recognition of images” . Lab work might involve training a CNN on a small dataset and visualizing sample predictions.

Module 8: Natural Language Processing (NLP) Basics (3 hours)

Description: Introduces ML techniques for text data. Topics include text preprocessing (tokenization, stop-word removal) and representing text as numbers (bag-of-words and word embeddings). Students implement a simple text classifier (e.g. sentiment analysis) using scikit-learn’s text vectorizers or using TensorFlow/PyTorch with an embedding layer. Theoretical focus is on how language can be modeled; as noted earlier, AI systems can handle “understanding language” when built with these tools . The module may also give a brief overview of advanced NLP (like recurrent networks or transformers) to show the direction of the field. Hands-on labs involve taking sample sentences and training a model to categorize them (e.g. spam vs. not spam).

Module 9: Capstone Project (10 hours)

Description: In this final, extensive project module, learners integrate everything they’ve learned. Students choose or are assigned a real-world problem (for example, predicting housing prices, classifying images, or analyzing text reviews) and go through the entire ML pipeline: data cleaning, exploratory analysis, model selection, and evaluation. The project emphasizes practical problem-solving: teams or individuals will “apply their theoretical knowledge to build models and even create an intelligent system”. Instructors provide guidance, but students carry out data collection/processing, code implementation, and interpret results. By completing the capstone, participants experience an end-to-end ML project, reinforcing skills from all prior modules. The outcome is a working model and a report or presentation summarizing the approach and findings.

Key Features

  Hands-on Lab Sessions: Interactive coding labs using Python (Jupyter notebooks) allow learners to apply theoretical concepts. Each module includes guided exercises where students write code to build models and analyze data in real time. Hands-on learning reinforces concepts by doing (for example, training on real datasets).

  Real-World Projects: Learners work on practical projects using real-world datasets (e.g. classification of images or text). These projects simulate actual industry tasks: “hands-on projects where students get practical experience by working on real-time data”. By building end-to-end models, participants bridge theory and practice.

  Assessments and Quizzes: Regular quizzes and mini-assignments at the end of each module test understanding of key concepts (e.g. regression vs classification, evaluation metrics). This ensures mastery of material and readiness for the next topic.

  Mentorship: The course provides guidance from experienced instructors who answer questions, review code, and give feedback. Small group discussions and Q&A sessions help beginners overcome hurdles.

  Certification: Upon successful completion, participants receive a certificate of training. Such certification (e.g. a formal certificate or badge) validates the new skills — for instance, a university ML course offers a printable certificate to those who pass.