Cart 0
Training
Course Information
Course Name
ML & AI: Machine Learning and Artificial Intelligence
Duration
5 Days
Overview
Audience Profile
Prerequisities
At Course Completion
Course Outline
Overview
This course focuses on the practical aspects of Machine Learning, Deep Learning and Artificial Intelligence. The objective is to make use of TensorFlow for various types of neural networks. The participants will build and train deep learning models.
Overview
Overview
This course focuses on the practical aspects of Machine Learning, Deep Learning and Artificial Intelligence. The objective is to make use of TensorFlow for various types of neural networks. The participants will build and train deep learning models.
Audience Profile
Audience Profile
The intended audience for this course:
• Engineers, analysts, and developers entering machine learning
• Aspiring data scientists and ML practitioners
• Technical professionals upskilling for AI-related roles
Prerequisities
Prerequisites
Participants are expected to have:
• Basic Python familiarity — variables, control flow, functions, and lists.
• Comfort with high-school statistics — mean, median, distributions, and correlation.
• A laptop with admin rights for installing Anaconda and Python libraries.
At Course Completion
At Course Completion
By the end of this program, participants will be able to:
• Explain core AI, ML, and Deep Learning concepts and how they relate.
• Apply Python with NumPy, Pandas, scikit-learn, and TensorFlow to real datasets.
• Build and evaluate supervised and unsupervised ML models end-to-end.
• Design, train, and tune deep neural networks, including CNNs, with Keras.
• Interpret model predictions using explainability tools such as SHAP and LIME.
Course Outline
Course Outline
Module 1: AI/ML Foundations & the Data-Science Workflow
• What AI, Machine Learning, Deep Learning, and Data Science actually mean
• How AI, ML, and DL relate — with practical examples for each
• Types of ML: supervised, unsupervised, semi-supervised, reinforcement
• Real-world ML use cases across industries
• The end-to-end ML project pipeline: problem → data → model → deploy
• Modern ML tools and libraries: scikit-learn, TensorFlow, Keras, XGBoost
• Setting up the environment: Anaconda, Jupyter, virtual environments
• Lab: explore a sample dataset in Jupyter and run a basic ML pipeline
Module 2: Python for ML — NumPy, Pandas, Visualization
• Python recap focused on ML: lists, dicts, comprehensions, functions
• NumPy arrays, vectorised operations, and broadcasting
• Pandas DataFrames: loading, indexing, filtering, joining, grouping
• Summary statistics and exploratory data analysis with Pandas
• Visualisation with matplotlib: line, bar, scatter, histogram
• Statistical visualisation with seaborn: distributions, pair plots, heatmaps
• Lab: end-to-end EDA on a structured dataset with a final notebook report
Module 3: Data Acquisition, Cleaning & Feature Engineering
• Loading data: CSV, Excel, JSON, and REST APIs with the requests library
• Web scraping basics with BeautifulSoup
• Handling missing values: deletion, imputation, indicators
• Detecting and handling outliers
• Encoding categorical variables: one-hot, label, target encoding
• Feature scaling: standardisation and normalisation
• Feature engineering: creating, transforming, and selecting features
• Train, validation, and test splits done correctly
• Lab: clean a messy real-world dataset and build a feature pipeline
Module 4: Supervised Learning — Classification & Regression
• Supervised learning: regression vs classification problem types
• Linear and logistic regression with scikit-learn
• Gradient descent and the role of loss functions
• Decision trees, k-nearest neighbours, and Naive Bayes
• Ensemble learning: Random Forest and Gradient Boosting
• Production-grade boosting with XGBoost
• The bias-variance trade-off and overfitting
• Regularisation: L1 and L2; early stopping
• Lab: build a classification model and a regression model on real data
Module 5: Unsupervised Learning — Clustering & Dimensionality Reduction
• Why unsupervised learning matters and where it applies
• K-Means clustering: algorithm, initialisation, choosing k
• Hierarchical clustering and DBSCAN for non-spherical clusters
• Cluster evaluation: silhouette score and inertia
• Principal Component Analysis for dimensionality reduction
• t-SNE and UMAP for visualising high-dimensional data
• Lab: customer segmentation with K-Means plus PCA visualization
Module 6: Model Evaluation, Tuning & Explainability
• Classification metrics: accuracy, precision, recall, F1, ROC-AUC, confusion matrix
• Regression metrics: MAE, MSE, RMSE, R-squared
• Cross-validation: k-fold, stratified, and time-series splits
• Hyperparameter tuning: grid search, random search, Bayesian optimisation
• Why explainability matters; introduction to XAI
• Global explanations: feature importance, permutation importance, SHAP
• Local explanations: LIME and SHAP for individual predictions
• Partial dependence and ICE plots for feature-target relationships
• Lab: tune a model and explain its predictions using SHAP and LIME
Module 7: Deep Learning with TensorFlow & Keras
• What Deep Learning is and how it differs from classical ML
• Neural network basics: perceptron, layers, weights, and activations
• Activation functions: ReLU, sigmoid, softmax, tanh
• Forward propagation, loss, and backpropagation in plain language
• TensorFlow 2 and the Keras-first API: Sequential and Functional models
• Training neural networks: optimisers, batch size, epochs, callbacks
• Regularisation in deep learning: dropout, batch normalisation, early stopping
• Convolutional Neural Networks: convolutions, pooling, and architectures
• Transfer learning with pre-trained Keras models
• Lab: build a fully-connected network, then a CNN for image classification
Module 8: Capstone Project
• Capstone goal: solve a realistic ML problem end-to-end
• Problem framing, success metrics, and dataset selection
• Data preparation, feature engineering, and model selection
• Training, tuning, and evaluating with the right metrics
• Explaining predictions using SHAP or LIME
• Teams of 2–3, mentored work session
• Final presentations: results, lessons learned, and next steps
