Course Content
KM-01: Overview of Artificial Intelligence
This module introduces learners to the fundamental concepts of Artificial Intelligence (AI) and its growing role in modern technology, business, and society. Learners will explore the evolution of AI, key definitions, and different types of artificial intelligence, as well as related fields such as machine learning, deep learning, neural networks, data science, automation, and robotics. The module also examines how AI is applied in real-world environments, including industries such as healthcare, finance, agriculture, manufacturing, and digital services. In addition, learners will understand the strategic advantages of AI in business, including automation, improved decision-making, and increased productivity. By the end of the module, learners will have a foundational understanding of AI technologies, their applications, and their impact on the Fourth Industrial Revolution (4IR). This knowledge prepares learners for further study and practical skills development within the Artificial Intelligence Software Developer qualification at NQF Level 4.
0/8
KM-02: Introduction to Mathematics and Statistics for Artificial Intelligence
This module introduces learners to the essential mathematical and statistical concepts required for understanding Artificial Intelligence, Machine Learning, Deep Learning, and Data Analytics. It provides foundational knowledge in areas such as basic mathematics, linear algebra, binary number systems, scientific notation, probability, and statistics. Learners will explore how mathematical principles are used to represent data, perform calculations, and analyze patterns in AI systems. The module also develops problem-solving skills through practical applications including coordinate systems, matrix operations, and probability models used in modern AI technologies.
0/25
KM-03: Analytical Thinking and Problem Solving
This module focuses on developing the learner’s ability to analyse problems logically and design structured solutions. Learners are introduced to analytical thinking techniques, critical thinking skills, and problem-solving methods used in artificial intelligence development. The module teaches how to break down complex problems, evaluate possible solutions, and apply structured reasoning when designing AI-based systems. By the end of the module, learners will understand how to approach real-world problems systematically and use analytical tools such as decision trees and critical thinking methods to support AI problem solving
0/7
KM-04: Data, Databases and Data Visualisation
This module introduces learners to the fundamental concepts of data, database systems, and data visualisation, which are essential components in modern artificial intelligence and data-driven technologies. The module focuses on helping learners understand how data is collected, processed, analysed, stored, and transformed into meaningful insights for decision-making. Learners begin by exploring the value of data and the role of data analysis, including how reliable data sources are identified and how raw data is refined by handling missing values, correcting misalignments, and eliminating irrelevant information. The module also explains common flaws and limitations in data collection, such as bias, omission, and errors that may affect the quality and reliability of data. The module then moves into practical data handling using spreadsheets, where learners study techniques for analysing and presenting data. This includes creating reports, sorting and filtering datasets, using pivot tables and dashboards, importing data from files and databases, and visualising results using charts and analytical tools. Learners are also introduced to databases and Structured Query Language (SQL), which allow large volumes of data to be stored, managed, and retrieved efficiently. In addition, the module explores data mining techniques used to identify patterns and relationships within datasets. Finally, the module highlights the importance of data visualisation and data security, teaching learners how to present information clearly using AI-assisted tools while ensuring that sensitive information is protected from misuse or unauthorized access. Overall, this module equips learners with the knowledge required to manage data effectively, perform analysis, create meaningful visualisations, and maintain data integrity and security, which are critical skills for professionals working in artificial intelligence, data science, and software development environments.
0/17
KM-05: Computing Theory
computational thinking. Programming is the process of writing instructions that tell a computer how to perform tasks. These instructions are written using programming languages such as Python, Java, or C++. In this module learners will develop an understanding of how computers interpret instructions, how algorithms are used to solve problems, and how basic programming structures work. The module also introduces the core principles of software development and provides an entry-level understanding of Python programming. By the end of the module learners will understand how software systems are designed, how algorithms are created to solve problems, and how programming languages are used to build modern digital solutions including artificial intelligence systems. The module covers the following key topics: Introduction to programming languages Introduction to algorithms Programming basics Solution development Introduction to Python These concepts provide the theoretical foundation needed before learners begin writing real programs in practical learning modules.
0/11
KM-06: Introduction to Artificial Intelligence, Machine Learning, Deep Learning
The main focus of the learning in this knowledge module is to build an understanding of the relationship between Artificial Intelligence, Machine Learning and Deep Learning, as well as the application of such systems to create a set of instructions to perform a programming task. Learners will explore how AI technologies are used across industries such as healthcare, finance, education, and automation. The module also introduces ethical considerations, responsible AI use, and the impact of AI on society and employment. By the end of this module, learners will understand how artificial intelligence systems work, the different types of AI technologies, and how these technologies are applied in modern software development environments.
0/3
KM-07: Artificial Intelligence Frameworks and Data Scraping
This module introduces learners to Artificial Intelligence frameworks and their role in developing intelligent systems. Learners will explore how frameworks such as TensorFlow, Keras, PyTorch and IBM Watson help developers design, train and deploy AI models efficiently. The module also introduces the concept of data scraping, explaining how AI technologies can be used to collect and extract information from websites. Learners will understand the tools, procedures, and legal considerations involved in web scraping and how this data can be used for analytics and decision-making. By the end of the module, learners will understand the structure of AI frameworks, their advantages, practical applications, and how AI techniques can be used to automate data extraction processes.
0/7
KM-08: Machine learning
The main focus of this knowledge module is to build an understanding of the relationship between Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning, as well as the application of machine learning to create a set of instructions that can perform programming tasks. This module introduces learners to the types of machine learning models, machine learning algorithm classifications, common machine learning algorithms, and the machine learning workflow process used to develop intelligent systems. Learners will also explore how machine learning can support business decision-making and improve business performance. The module further explains how machine learning systems use data, features, and labels to identify patterns, make predictions, and automate tasks. By understanding these concepts, learners will gain the foundational knowledge required to work with machine learning technologies and apply them in real-world applications and business environments.
0/11
KM-09: Deep Learning (DL)
This module introduces learners to the concept of Deep Learning, an advanced area of Artificial Intelligence that builds on Machine Learning techniques to create intelligent systems capable of learning complex patterns from large datasets. The module focuses on understanding the relationship between Artificial Intelligence (AI), Machine Learning (ML), and Deep Learning (DL) and how deep learning technologies are used to develop intelligent applications. Learners will explore how neural networks are structured and how they function, including the roles of input layers, hidden layers, and output layers in deep learning systems. The module also introduces different neural network architectures such as convolutional neural networks, recurrent neural networks, and recursive neural networks, which are widely used in fields such as computer vision, natural language processing, and speech recognition. In addition, the module covers activation functions used in deep learning models, including functions such as Sigmoid, Tanh, Softmax, and ReLU. Learners will also study how deep learning networks are built, trained, and tuned to improve performance. These concepts help developers design more accurate and efficient models for solving complex computational problems. The module further introduces advanced Python concepts for deep learning, including decorators, context managers, exception handling, and Python package management. These programming techniques are important for developing scalable deep learning applications. Finally, learners will explore TensorFlow and Keras, two of the most widely used frameworks for deep learning development. These tools allow developers to build, train, and deploy neural networks efficiently using modern machine learning libraries and APIs. By the end of this module, learners will understand the core concepts of deep learning, neural network architecture, advanced Python programming for AI development, and the use of TensorFlow and Keras to build deep learning models.
0/7
KM-10: Introduction to Governance, Legislation and Ethics
This module introduces learners to the principles of governance, legislation, ethics, workplace security, and business practices that influence organisations and employees. The module focuses on understanding how legal frameworks and ethical standards guide behaviour in the workplace and ensure accountability, transparency, and responsible decision-making. Learners will explore important workplace legislation such as the Labour Relations Act (LRA), the Protection of Personal Information Act (POPIA), and other regulatory frameworks that affect employees and employers. The module also introduces key ethical principles, including professional conduct, fairness, honesty, and accountability in professional environments. In addition, the module examines workplace security, performance management, business planning, and costing concepts that influence organisational efficiency and sustainability. By the end of the module, learners will understand how governance, ethics, legislation, and management practices contribute to a responsible and productive workplace environment.
0/19
KM-11: Fundamentals of Design Thinking and Innovation
This module introduces learners to the principles of design thinking, creativity, and innovation in the workplace. It focuses on solving problems using a human-centered approach, where user needs are prioritised through observation, empathy, and iterative development. Learners will explore key concepts such as design thinking methodology, creativity, innovation types, and application in real-world environments, including software development and business. The module also highlights how organisations use design thinking to improve products, processes, and services while fostering innovation. By the end of this module, learners will understand how to apply design thinking to solve complex problems and drive innovation effectively in the workplace.
0/15
KM-12: Fundamentals of Research and Information Analysis
This module focuses on developing an understanding of research principles, information gathering, and data analysis techniques. It equips learners with the ability to collect, evaluate, interpret, and apply information effectively in problem-solving and decision-making contexts
0/6
Artificial Intelligence Software Developer

TensorFlow and Keras for Deep Learning (KM-09-KT03)

Lesson Overview

Deep learning applications often require powerful frameworks that simplify the process of building, training, and deploying neural networks. Two of the most widely used frameworks for deep learning development are TensorFlow and Keras.

TensorFlow is an open-source artificial intelligence library that allows developers to create machine learning and deep learning models using data flow graphs and tensors. It provides powerful tools for designing large-scale neural networks and performing complex computations.

Keras is a high-level API integrated into TensorFlow that simplifies the development of neural networks. It provides an intuitive interface that allows developers to quickly design and train deep learning models.

This lesson explores the core concepts of TensorFlow, tensors, control structures, Keras models, sequential models, configuration layers, and data loading processes used in deep learning development.

Learning Outcomes

By the end of this lesson, learners should be able to:

  • Understand the concept of TensorFlow and its purpose
  • Explain what a tensor is in TensorFlow
  • Identify the types of control structures in programming
  • Understand how Keras integrates with TensorFlow
  • Identify the types of sequential models used in Keras
  • Explain the purpose of configuration layers
  • Understand the concept of data loading in machine learning systems

1. Introduction to TensorFlow

TensorFlow is an open-source artificial intelligence and machine learning framework developed by Google. It is widely used for building and training deep learning models.

TensorFlow uses data flow graphs to represent computations. In these graphs, nodes represent mathematical operations, while edges represent data arrays known as tensors.

TensorFlow is used for several AI tasks including:

  • Classification
  • Prediction
  • Pattern recognition
  • Natural language processing
  • Computer vision
  • Recommendation systems

TensorFlow enables developers to build large-scale neural networks with multiple layers, making it one of the most powerful tools for deep learning development.

2. Understanding Tensors in TensorFlow

A tensor is the fundamental data structure used in TensorFlow.

A tensor is a generalization of vectors and matrices to higher dimensions. It represents multidimensional arrays that store numerical data used in machine learning models.

Examples of tensors include:

  • A scalar (single value)
  • A vector (one-dimensional array)
  • A matrix (two-dimensional array)
  • Higher-dimensional arrays used in deep learning computations

When writing TensorFlow programs, the primary object manipulated by the system is the TensorFlow tensor (tf.Tensor).

Tensors allow deep learning systems to perform complex numerical operations required for training neural networks.

3. Control Structures in Programming

A control structure is a programming construct that determines how a program executes instructions based on conditions or loops.

Control structures allow programs to make decisions and repeat operations when necessary.

The main types of control structures include:

  1. Sequence logic – instructions are executed in order
  2. Selection logic – decisions are made using conditions such as if statements
  3. Iteration logic – instructions are repeated using loops

These control structures are essential for building machine learning algorithms and controlling the flow of computations in a program.

4. TensorFlow APIs

TensorFlow provides several Application Programming Interfaces (APIs) that allow developers to build machine learning models.

An API is a set of functions and tools that allow developers to interact with software libraries.

TensorFlow APIs are available in multiple programming languages, but the Python API is the most complete and widely used. Python allows developers to easily build deep learning models and integrate them into applications.

TensorFlow APIs allow developers to:

  • Construct machine learning models
  • Train neural networks
  • Evaluate model performance
  • Deploy models into production systems

5. Keras: High-Level API for Deep Learning

Keras is a high-level API integrated into TensorFlow that simplifies the development of deep learning models.

Keras provides easy-to-use building blocks that allow developers to create neural networks quickly and efficiently.

Key advantages of Keras include:

  • Simple and user-friendly interface
  • Rapid prototyping of machine learning models
  • Integration with TensorFlow backend
  • Support for complex neural network architectures

Keras allows developers to focus on designing models rather than writing low-level code.

6. Sequential Models in Keras

Keras provides two primary types of models used to build neural networks:

  1. Sequential Model
  2. Functional Model

The Sequential Model is the simplest type of model in Keras. It allows developers to create neural networks by stacking layers sequentially from input to output.

Sequential models are commonly used for simple neural networks where layers are arranged in a linear sequence.

The Functional Model is more flexible and allows developers to build complex architectures with multiple inputs, outputs, or branching layers.

These models allow developers to design advanced deep learning architectures for complex applications.

7. Configuration Layers

A configuration layer is a customized set of metadata used to configure variables and data sources for machine learning models or reports.

Configuration layers help organize and manage the settings required to run machine learning models efficiently.

The configuration information is typically defined by system designers and stored in templates that are used during system execution.

This structure helps simplify the management of complex machine learning systems.

8. Data Loading in Machine Learning

Data loading is the process of transferring data from a source into a system where it can be processed or analyzed.

In machine learning systems, data loading involves copying data from sources such as:

  • Files
  • Databases
  • External applications
  • Cloud storage systems

The data is then loaded into memory or storage systems where machine learning algorithms can process it.

Data loading is an important step in the machine learning pipeline because the quality and structure of the data influence the performance of the model.

Efficient data loading techniques help improve the speed and performance of machine learning systems.

Lesson Summary

TensorFlow and Keras are powerful frameworks used for developing deep learning models. TensorFlow provides the computational infrastructure required for building neural networks using tensors and data flow graphs.

Tensors serve as the core data structures used in deep learning computations. Programming control structures such as sequence, selection, and iteration help manage the flow of machine learning algorithms.

Keras simplifies deep learning development by providing a high-level API that allows developers to quickly build neural networks using sequential or functional models.

Additional concepts such as configuration layers and data loading help organize machine learning systems and ensure that data is properly processed for training and prediction tasks.

Understanding these tools and concepts enables developers to build scalable and efficient deep learning applications.

 

Scroll to Top