II.            Preprocess and analyze datasets, handling tasks like feature engineering, normalization, and splitting data for training/validation, in order to feed data appropriately into ML models.

   III.            Implement and train various ML models using Python (from scratch for simple algorithms and using libraries for more complex ones) and tune their hyperparameters to improve performance.

   IV.            Evaluate model performance using appropriate metrics (accuracy, precision/recall, RMSE, etc.) and techniques like cross-validation, and interpret the results to make informed decisions about model selection.

      V.            Compare model performance and complexity to choose a suitable algorithm for a given problem, justifying the choice based on data characteristics and use-case requirements.

   VI.            Understand basic theoretical concepts such as overfitting vs. underfitting and how techniques like regularization or ensemble methods help improve generalization. This includes being able to articulate the bias-variance trade-off in the context of model complexity.

4.      Deep Learning

This course builds upon Machine Learning to focus on deep neural networks and modern representation-learning techniques. Students explore the concepts and architectures that have driven recent breakthroughs in AI. The course begins with an introduction to neural network fundamentals: perceptrons, activation functions, multilayer feed-forward networks, and backpropagation for training. It then progresses to advanced architectures such as Convolutional Neural Networks (CNNs) for image data, Recurrent Neural Networks (RNNs) and sequence models (including LSTM and GRU networks) for sequential data, and Transformer architectures for sequence transduction and attention mechanisms (the basis of modern NLP). Topics like autoencoders, generative models (e.g. GANs and variational autoencoders), and deep reinforcement learning basics may be introduced as well. Practical skills are heavily emphasized: students will implement neural networks using frameworks like TensorFlow or PyTorch, and train models on GPUs for tasks such as image classification (e.g. recognizing objects in images) and text classification or generation. The course covers training techniques (optimization algorithms beyond basic SGD, like Adam; initialization; batch normalization), as well as issues of tuning deep models and preventing overfitting (dropout, regularization). Recent research papers may be discussed to highlight state-of-the-art developments (for example, breakthroughs in generative AI like GPT for language or diffusion models for image generation). By the end, students will be able to design, train, and deploy deep neural network models for a variety of data types.

Learning Outcomes

 Upon completion, students will be able to:

        I.            Explain the architecture of neural networks, including how neurons are organized into layers and how backpropagation updates model weights through gradient descent.

      II.            Construct and train deep learning models for different tasks: CNNs for vision tasks (e.g. image recognition), RNN/Transformer models for sequence data (e.g. language modeling), and others, using industry-standard libraries.

   III.            Configure training processes effectively – selecting appropriate loss functions and optimization algorithms, tuning hyperparameters like learning rate and network depth, and using regularization techniques to improve generalization.

   IV.            Analyze training outcomes via learning curves and performance metrics, and troubleshoot common issues in deep learning (such as vanishing/exploding gradients, overfitting, or data imbalance).

      V.            Understand and implement basic generative modeling techniques (e.g. designing an autoencoder or a simple GAN) and discuss their applications in creating new data (such as image synthesis or text generation).

   VI.            Stay abreast of current research trends in deep learning, demonstrating the ability to read a contemporary research paper (for instance on a new model or optimization method) and summarize its contributions. (This “learning to learn” skill is cultivated to help graduates keep up with AI advances).

5.      Natural Language Processing

This course provides a comprehensive introduction to Natural Language Processing (NLP) – the branch of AI concerned with understanding and generating human language. It covers both foundational linguistic concepts and the modern, machine learning-based approaches to NLP. Major topics include: text preprocessing (tokenization, stemming, embeddings), language models (from n-grams to neural language models), syntactic parsing (POS tagging, context-free grammars, dependency parsing), and semantic processing (word sense disambiguation, ontology use). The course places significant emphasis on deep learning for NLP, including word embeddings (word2vec, GloVe) and advanced architectures like Recurrent Neural Networks and Transformer-based models (such as BERT and GPT) which enable tasks like named entity recognition, machine translation, question answering, sentiment analysis, and chatbot dialogue systems. Students will engage in lab-style assignments to build NLP pipelines – for example, constructing a sentiment analyzer for social media text or a simple machine translation system – using Python NLP libraries (NLTK, spaCy) and deep learning frameworks. The interplay between NLP and cognitive science is touched upon, and ethical considerations (such as bias in language models or responsible use of generative text like ChatGPT) are discussed. By completing this course, students gain the ability to develop systems that process and derive meaning from natural language data.

Learning Outcomes

 Upon completion, students will be able to:

         I.            Apply linguistic preprocessing techniques to raw text (tokenizing sentences/words, normalizing text, extracting features) as a necessary step before feeding data into NLP models.

  II.            Understand and implement core NLP algorithms for several tasks: e.g. a part-of-speech tagger (using sequence labeling methods), a basic parser for sentence structure, and a named entity recognition system.

      III.            Utilize machine learning and deep learning methods in NLP, including training and using recurrent or transformer models for tasks like language modeling, text classification, and translation.

   IV.            For instance, students might train a simple transformer-based classifier for news topics or fine-tune a pre-trained BERT model for an NLP task.

   V.            Represent text data through embeddings (learning vector representations for words or sentences) and explain how these representations capture semantic meaning (e.g. similar words having vectors close together).

      VI.            Evaluate NLP models with appropriate metrics (BLEU scores for translation, F1 for named entity recognition, etc.) and error analysis, and iterate to improve their performance.

   VII.            Appreciate the ethical and societal implications of NLP technologies – including issues of bias in language models or misinformation – and articulate strategies to mitigate harm (this ties in with the Ethics course).

 VIII.            (Stretch goal) Experiment with a large language model (LLM) or generative model and understand at a high level how such models are enabling the current “Generative AI” trend, as well as limitations like hallucination or large compute requirements.

6.      Computer Vision

This course covers the principles and techniques of Computer Vision, enabling computers to interpret and understand visual information from images or videos. It starts with fundamentals of image processing and classic vision algorithms: image formation and representation, color spaces, filtering and convolution, edge detection, feature extraction (SIFT, SURF), and segmentation methods. Students learn about geometric computer vision concepts such as camera models, stereoscopic vision, and 3D reconstruction basics. Building on these fundamentals, the course then focuses on modern deep learning approaches to vision, including Convolutional Neural Networks (CNNs) for image classification and object detection (architectures like ResNet, YOLO, or Mask R-CNN are discussed). Topics like image classification, object recognition, face detection, motion and tracking in video, and perhaps an introduction to image generation (GANs) are included. Throughout the course, students gain practical experience via programming projects: for example, implementing an edge detector, building an image classifier for a set of images, or using OpenCV and PyTorch to create a real-time object detection system. The course also surveys applications of computer vision in industry (e.g. in autonomous vehicles, medical image analysis, or retail analytics). By combining foundational vision knowledge with hands-on neural network training for images, students acquire the skills to build and deploy vision-based AI solutions.

Learning Outcomes

 Upon completion, students will be able to:

        I.            Understand the image formation process and basic image properties (pixels, color channels), and perform fundamental image processing operations (smoothing, edge detection, segmentation) to extract useful features.

      II.            Implement and apply feature detection and description algorithms (e.g. edge detectors like Canny, corner detectors, SIFT features) to identify salient points or regions in images for further analysis.

   III.            Design and train deep CNN models for vision tasks – such as classifying images into categories or detecting objects within images – using frameworks like TensorFlow/PyTorch. This includes understanding how convolution, pooling, and deep layers enable hierarchical feature learning in vision.

   IV.            Evaluate the performance of vision systems using metrics like accuracy, Intersection-over-Union (for object detection), etc., and improve models by techniques such as data augmentation or transfer learning from pre-trained models.

      V.            Solve a real-world computer vision problem end-to-end (for instance, build a prototype that takes webcam input and identifies a set of objects), demonstrating integration of vision algorithms into an application.

   VI.            Discuss advanced or emerging topics in vision (e.g. the role of vision in autonomous systems or video analytics), showing awareness of how classical methods and deep learning combine in state-of-the-art solutions.

 VII.            Recognize the challenges and limitations of computer vision systems (lighting variation, viewpoint variation, need for large training data, etc.) and propose methods to address these in practical deployments.

      II.            Containerize and deploy an AI model as a web service or API, using technologies like Docker (for packaging code and model) and Flask/FastAPI or cloud functions (to serve predictions). They will understand how to expose model predictions to other software components or end-users securely.

   III.            Utilize cloud computing resources for AI: e.g. launching and configuring a cloud VM or using a managed service to train a model on GPU/TPU, then deploying the model with auto-scaling. They will be familiar with one major cloud provider’s AI ecosystem (AWS, Azure, or GCP) in the context of deploying machine learning.

   IV.            Implement monitoring and logging for AI systems in production – capturing metrics like latency, throughput, and prediction accuracy over time – and set up alerts or triggers for when models degrade or data drifts.

      V.            Explain and apply DevOps principles to ML (MLOps), including version control for datasets and models, continuous integration (automated testing of model code), and continuous deployment (automated rollout of model updates).

        VI.            Address important practical considerations such as scalability (understanding how to handle increasing loads or larger data via distributed computing), privacy and security (ensuring data is handled securely, models are protected from abuse, and compliance with regulations like GDPR when deploying AI), and cost optimization (balancing computational cost with performance needs in cloud deployments).

      VII.            Work effectively in a multidisciplinary team (data scientists, software engineers, domain experts) to deliver an AI system, communicating requirements and results – this mimics the real-world scenario of deploying AI in a business or organizational context.

8.      Ethics and Policy in Artificial Intelligence

This core course examines the ethical, legal, and policy implications of AI technologies. As AI systems become increasingly integrated into society, professionals must be equipped to address questions of fairness, accountability, transparency, and social impact.

The course is structured in three modules: (1) Ethics of AI – covering fundamental ethical frameworks (deontology, utilitarianism, etc.) as applied to AI, and key issues such as algorithmic bias/discrimination, privacy and surveillance, autonomy (e.g. lethal autonomous weapons) and the implications of AI on labor and the economy; (2) AI Policy and Governance – exploring how governments and institutions are responding to AI (national AI strategies, emerging regulations like the EU AI Act, guidelines for AI ethics), including topics like data protection law, intellectual property issues for AI-generated content, and governance of AI in specific sectors (healthcare, finance, criminal justice); (3) Responsible AI Practice – focusing on practical approaches to developing and deploying AI ethically, such as fairness-aware machine learning, explainability techniques, model auditing, and the role of AI ethics committees or review boards. Case studies (e.g. facial recognition in law enforcement, use of AI in hiring decisions, autonomous vehicle accidents, large language models like ChatGPT generating misinformation) are analyzed to illustrate the real-world dilemmas and decision-making processes. Students will engage in discussions, write reflective essays or policy memos, and possibly collaborate on creating an “AI ethics impact assessment” for a hypothetical AI product. The course may feature guest lectures from experts in AI law or industry ethics panels. By the end, students gain a nuanced understanding of how to balance innovation with responsibility and are prepared to make informed judgments about AI use in their careers.

Learning Outcomes

 Upon completion, students will be able to:

       I.            Identify and analyze ethical issues in AI scenarios, applying principles of ethical reasoning to questions like: Is an AI decision fair? Is it transparent? Does it respect user privacy and autonomy? 

   II.            Evaluate the societal impact of AI systems, including potential benefits and harms to different stakeholder groups, and articulate these in the context of case studies (e.g. evaluating an AI system’s impact on employment or on marginalized communities).

   III.            Demonstrate knowledge of current and emerging AI regulations and policies – for instance, students should be able to summarize key points of guidelines like the OECD AI Principles or provisions of the EU’s draft AI Act, and understand how these affect AI deployment in practice.

   IV.            Implement basic practices of ethical AI development, such as conducting bias audits on a dataset/model (and interpreting bias metrics), improving model transparency (through documentation and explainable AI techniques), and considering accessibility and inclusivity in design.

      V.            Understand the concept of AI governance within organizations, including the roles of ethics committees, frameworks for AI accountability, and how companies can self-regulate through standards (like IEEE’s AI ethics standards or NIST’s AI risk management framework).

   VI.            Formulate a policy or management recommendation concerning an AI issue (for example, writing a memo on whether a city should adopt predictive policing software, weighing ethical concerns and evidence). This demonstrates the ability to bridge technical and policy perspectives, a skill valuable in leadership roles.

 VII.            Appreciate the long-term and global challenges of AI (such as AI’s effect on democracy, the environment, or the future of work) and discuss them critically – aligning with the programme’s goal of producing responsible leaders in AI.

9.      AI Strategy and Business Applications

This capstone-core course connects the technical facets of AI with business strategy and industry applications. It is tailored for a business school context, focusing on how organizations can generate value from AI and how to lead AI-driven innovation. The course covers frameworks for developing an AI strategy aligned with business objectives, including identifying suitable AI use cases, conducting cost-benefit and ROI analysis for AI projects, and managing data as a strategic asset. Students survey AI applications across key industries – e.g. predictive analytics in finance, personalization in marketing, diagnostic AI in healthcare, manufacturing automation, supply chain optimization, etc. – analyzing what has made certain AI implementations successful or why some have failed. Concepts of AI product management are introduced: how to take an AI idea from concept to deployment, including gathering requirements, prototyping, testing with users, and iterating. The course also deals with change management and organizational readiness for AI: addressing how to build AI talent teams, foster data-driven culture, and manage the ethical and workforce implications of introducing AI into an enterprise. Students will learn about governance and oversight in a corporate context (complementing the Policy course, but here from a managerial perspective), such as establishing AI ethics guidelines within a company and ensuring compliance. Another component is understanding the landscape of AI solutions and vendors (e.g. evaluating when to use cloud AI services vs. custom solutions). Guest speakers from industry (AI project managers, consultants, or executives) may share real-world insights. Assessment may involve students developing an AI strategy proposal for an existing company or a startup idea, including business case, implementation roadmap, and risk analysis. By blending technical understanding with business insight, this course prepares students to take on roles interfacing between technical teams and business leadership.

Learning Outcomes

 Upon completion, students will be able to:

         I.            Identify high-impact AI opportunities in a business context by analyzing business processes and challenges, and recognizing where AI techniques (like predictive modeling, computer vision, NLP, etc.) could provide innovative solutions or efficiency gains.

      II.            Develop a comprehensive AI project proposal or business plan that includes problem definition, solution approach (which AI technologies to use), resource estimation, cost-benefit analysis, and KPIs for success. Students will practice articulating the business value of the technical solution in language suitable for executives.

   III.            Understand how to integrate AI systems into existing operations, including considerations for IT infrastructure, data governance, and cross-department collaboration. They will be able to outline how an AI pilot project can be scaled up to organization-wide deployment.

   IV.            Discuss and plan for the organizational implications of AI adoption: for example, training/upskilling employees, redefining job roles (rather than simply automating tasks, how to use AI to augment human workers productively), and addressing employee or customer concerns about AI (trust, transparency).

      V.            Apply principles of AI product management, such as user-centered design for AI (ensuring the AI tool is usable and solves the right problem), iterative development with user feedback, and maintenance planning (updates as data changes or new features are needed).

   VI.            Evaluate commercial AI solutions and platforms – given a problem, decide whether to build a custom model in-house or use an existing API/service, considering factors like time-to-market, cost, data sensitivity, and competitive advantage.

 VII.            Present and communicate AI-driven initiatives effectively to non-technical stakeholders, demonstrating the ability to be an “AI translator” between data scientists and business leaders – a skill strongly sought after in modern enterprises.

VIII.            This includes being conversant in both technical terms and business strategy terms, thereby leading cross-functional teams in executing AI projects.

Course Learning Outcomes

Upon completing TBS’s AI & Python program, you will be able to:

• Confidently Code in Python
  Write clear, efficient, and well-structured Python code that meets professional standards and supports quick adaptation to emerging AI technologies.
• Integrate AI Models into Real Projects
  Seamlessly incorporate machine learning and deep learning algorithms into Python applications, with a focus on delivering tangible business value.
• Automate and Streamline Complex Tasks
  Harness Python’s rich ecosystem of libraries to simplify data cleaning, accelerate predictive analytics, and optimize day-to-day operations.
• Visualize and Communicate Insights
  Present analytical findings effectively using Python’s visualization tools, transforming dense data into compelling narratives for stakeholders.
• Strategize AI Deployments
  Plan, execute, and manage AI initiatives that align with organizational objectives—becoming the catalyst for innovation in any team you join.

Learning Outcomes

 By the end of the course, students will be able to:

         I.            Model decision-making problems using Markov Decision Processes.

       II.            Implement dynamic programming algorithms for solving MDPs.

    III.            Apply model-free reinforcement learning techniques like Q-learning and SARSA.

     IV.            Use deep reinforcement learning methods such as DQN and policy gradients.

       V.            Train agents using simulation platforms like OpenAI Gym.

     VI.            Analyse and optimise agent behaviour in real-world and synthetic environments.

   VII.           ​ Evaluate RL strategies in applications such as games, robotics, and personalised systems.

Learning Outcomes

 By the end of the course, students will be able to:

          I.            Understand and apply the fundamentals of robotic kinematics.

       II.            Integrate perception systems using LIDAR and camera-based vision.

    III.            Perform sensor fusion to enhance robot perception and localisation.

     IV.            Implement SLAM and obstacle avoidance algorithms for autonomous navigation.

       V.            Design decision-making systems for autonomous robots.

     VI.            Apply reinforcement learning or control theory in robotic applications.

   VII.            Use ROS to program robots for real-world or simulated tasks.

VIII.            Evaluate the performance of autonomous systems in dynamic environments.

Learning Outcomes

 By the end of the course, students will be able to:

          I.            Design and implement GANs, VAEs, and other generative models.

       II.            Build and analyse Transformer-based architectures for text and image tasks.

    III.            Understand and apply diffusion models in generative image creation.

     IV.            Explore and experiment with multimodal models combining vision and language.

       V.            Apply model compression techniques for efficient deployment of large AI models.

     VI.            Analyse the challenges and strategies in large-scale model training.

   VII.            Critically read, replicate, and extend recent research in advanced deep learning.

VIII.            Develop a project that demonstrates mastery of cutting-edge generative AI techniques.

       II.            Differentiate between SQL and NoSQL databases for storing and retrieving AI-relevant data.

    III.            Design scalable cloud-based data pipelines for AI model training and deployment.

     IV.            Optimise data handling processes for efficiency and performance on large datasets.

       V.            Integrate cloud platforms into end-to-end AI workflows.

     VI.            Build and test full AI pipelines from data ingestion to model output.

   VII.            Address scalability and resource management challenges in big data AI projects.

       II.            Build predictive models using patient health data for diagnosis or risk assessment.

    III.            Use NLP techniques to extract insights from clinical text data.

     IV.            Navigate ethical and legal considerations in healthcare AI applications.

       V.            Develop AI-based tools tailored to healthcare environments.

     VI.            Work with real-world de-identified clinical datasets.

   VII.            Understand regulatory pathways for AI approval in medical settings.

VIII.            Evaluate the impact and limitations of AI in clinical decision-making and healthcare delivery.

The Research Lab Practicum allows students to join university research teams and contribute to ongoing AI projects. Ideal for those interested in academic or research careers, students may work on tasks like developing algorithms, analysing datasets, or co-authoring demos and papers. This path often leads to research publications or future doctoral study.

Regardless of the path chosen, students complete a Practicum Report and Presentation, demonstrating the AI solution built, outcomes achieved, and ethical or business considerations addressed. Graded on a pass/fail basis, the practicum builds job-ready skills and often leads directly to job offers or strong portfolio projects, ensuring graduates leave the programme with both theoretical depth and proven practical competence.

The Microsoft Certified: Azure AI Engineer Associate certification is covered through a combination of foundational ML courses and the AI Systems Design and AI Strategy courses. These teach students how to build, deploy, and optimise AI models using Azure services. Access to Azure for Education credits allows students to gain practical experience with Azure AI tools, making them well-prepared for scenario-based certification questions.

Through this optional certification alignment, students graduate not only with academic credentials but also with the practical skills and readiness to pursue globally recognised AI certifications.