SAQA All qualifications and part qualifications registered on the National Qualifications Framework are public property. Thus the only payment that can be made for them is for service and reproduction. It is illegal to sell this material for profit. If the material is reproduced or quoted, the South African Qualifications Authority (SAQA) should be acknowledged as the source.
SOUTH AFRICAN QUALIFICATIONS AUTHORITY 
REGISTERED QUALIFICATION: 

Occupational Certificate: Artificial Intelligence Software Developer 
SAQA QUAL ID QUALIFICATION TITLE
118792  Occupational Certificate: Artificial Intelligence Software Developer 
ORIGINATOR
Development Quality Partner-MICT SETA 
PRIMARY OR DELEGATED QUALITY ASSURANCE FUNCTIONARY NQF SUB-FRAMEWORK
-   OQSF - Occupational Qualifications Sub-framework 
QUALIFICATION TYPE FIELD SUBFIELD
Occupational Certificate  Field 10 - Physical, Mathematical, Computer and Life Sciences  Information Technology and Computer Sciences 
ABET BAND MINIMUM CREDITS PRE-2009 NQF LEVEL NQF LEVEL QUAL CLASS
Undefined  209  Not Applicable  NQF Level 05  Regular-ELOAC 
REGISTRATION STATUS SAQA DECISION NUMBER REGISTRATION START DATE REGISTRATION END DATE
Registered  EXCO 0522/24  2022-02-03  2025-12-31 
LAST DATE FOR ENROLMENT LAST DATE FOR ACHIEVEMENT
2026-12-31   2029-12-31  

In all of the tables in this document, both the pre-2009 NQF Level and the NQF Level is shown. In the text (purpose statements, qualification rules, etc), any references to NQF Levels are to the pre-2009 levels unless specifically stated otherwise.  

This qualification does not replace any other qualification and is not replaced by any other qualification. 

PURPOSE AND RATIONALE OF THE QUALIFICATION 
Purpose:
The purpose of the Occupational Certificate: Artificial Intelligence Software Developer is to prepare a learner to operate as an Artificial Intelligence Software Developer

Artificial Intelligence Software Developers build Artificial Intelligence (AI) functionality into software applications through integrating and implementing AI algorithms and logic into the deliverables of an Information Technology (IT) project. Developers teach the machine to solve problems the way a human would through the use of programming. They create, test and deploy code. These developers also assist in converting machine learning Application Programming Interface (APIs) so that other applications can use them.

A qualified learner will be able to:
  • Interpret solution design documentation and develop AI solution.
  • Train the AI model through a machine learning process and test the performance to ensure that model accuracy is strictly maintained within the selection framework.
  • Deploy the AI solution and maintain the solution to ensure model accuracy is strictly maintained.

    Rationale:
    This qualification has been developed in response to the report of the Presidential Commission on the 4th Industrial Revolution (4IR). This report forefronts human capital and the future of work and refers to growing skills instability. The extent of 4IR today and its impact on businesses and the economy, is unparalleled. This implies that companies need to urgently prepare as AI will shape the future of our world more powerfully than any other innovation this century.

    Research findings indicate that businesses are expected to hire more technology and automation professionals in the future, pointing to the need for well-qualified developers in the AI field. The most sought-after areas of expertise include artificial intelligence, digital customer experience, internet of things and the cloud.

    AI seems to be revolutionary in its ability to transform the way organisations operate, but without a workforce skilled in technology, any effort by businesses to embrace the adoption of AI initiatives will fail. This qualification focusses on establishing a firm understanding of AI technology, its applications and its use cases.

    A Masters degree in Machine Learning and Artificial Intelligence exists. This Occupational Certificate: Artificial Intelligence Software Developer is thus unique and filling a glaring gap at the entry level of this specialist field and career path. Software developers create the applications or systems that run on a computer or another device. AI software developers build AI functionality into software applications. The role is generally focused on integrating and implementing AI algorithms and logic into the deliverable. AI, machine learning, neural networks and deep learning is a unique field, using specialist tools to create intelligence that mimics human interaction. AI developers use machine learning to create intelligence.

    The IT sector and the economy will benefit from AI as it can dramatically improve the efficiencies of our workplaces, augmenting the work humans can do and reducing the consequences of human error. When AI takes over repetitive or dangerous tasks, it frees up the human workforce and enables people to use their creativity for verification, validity, security, control and so forth.

    There are many ways in which AI will benefit society: AI is facilitating major advances in healthcare, and it improves customer service and the human experience. In short, AI has revolutionised the ease and dynamics with which humans interact with their environment.

    The typical learners identified for this qualification are school leavers, qualified learners from TVET colleges, new entrants into the sector and existing employees who have experience in this field, but without formal recognition of skills and competencies. Professionals who want to augment their careers may also access this qualification. The qualification is structured to enable individuals without prior learning in the field of information technology to acquire the knowledge and skills to seamlessly enter the field of work.

    No professional registration or licencing is expected for AI Software Developers to seek employment in the sector. AI Software Developers can be employed as AI Researchers, Machine Learning Engineers, Machine Learning Researchers, AI Architects, AI Engineers, AI Technicians, AI Developers, Business Intelligence (BI) Developers and working with engineering teams in the field of intelligent robotics. 
    		•

    LEARNING ASSUMED TO BE IN PLACE AND RECOGNITION OF PRIOR LEARNING 
    Recognition of Prior Learning (RPL):
    RPL for Access to the External Integrated Summative Assessment.
    Accredited providers and approved workplaces must apply the internal assessment criteria specified in the related curriculum document to establish and confirm prior learning. Accredited providers and workplaces must confirm prior learning by issuing a statement of result.

    RPL for Access to the Qualification:
  • Learners will gain access to the qualification through RPL for Access as provided for in the QCTO RPL Policy. RPL for access is conducted by accredited education institution, skills development provider or is workplace accredited to offer that specific qualification/part qualification.
  • Learners who have acquired competencies of the modules of a qualification or part qualification will be credited for modules through RPL.

    RPL for access to the external integrated summative assessment:
    Accredited providers and approved workplaces must apply the internal assessment criteria specified in the related curriculum document to establish and confirm prior learning. Accredited providers and workplaces must confirm prior learning by issuing a statement of result.

    Entry Requirements:
    The minimum entry requirement for this qualification is:
  • NQF Level 4 qualification. 
    		•

    RECOGNISE PREVIOUS LEARNING? 
    Y 

    QUALIFICATION RULES 
    This qualification is made up of compulsory Knowledge, Practical Skill and Work Experience Modules:
    Knowledge Modules
  • 251201-002-00-KM-01 Overview of Artificial Intelligence, Level 4, 2 Credits.
  • 251201-002-00-KM-02 Introduction to Mathematics and Statistics, Level 4, 10 Credits.
  • 251201-002-00-KM-03 Analytical Thinking and Problem Solving, Level 4, 3 Credits.
  • 251201-002-00-KM-04 Data, Databases and Data Visualisation, Level 4, 8 Credits.
  • 251201-002-00-KM-05 Computing Theory, Level 4, 8 Credits.
  • 251201-002-00-KM-06 Introduction to Artificial Intelligence, Machine Learning, Deep Learning, Level 4, 5 Credits.
  • 251201-002-00-KM-10 Introduction to Governance, Legislation and Ethics, Level 4, 1 Credit.
  • 251201-002-00-KM-11 Fundamentals of Design Thinking and Innovation, Level 4, 1 Credit.
  • 251201-002-00-KM-12 4IR and Future Skills, Level 4, 4 Credits.
  • 251201-002-00-KM-07 Artificial Intelligence, Level 5, 12 Credits.
  • 251201-002-00-KM-08 Machine Learning, Level 5, 16 Credits.
  • 251201-002-00-KM-09 Deep Learning, Level 5, 16 Credits.

    Total number of credits for Knowledge Modules: 86

    Practical Skill Modules
  • 251201-002-00-PM-01 Mathematics and Statistics for Programming, Level 4, 8 Credits.
  • 251201-002-00-PM-02 Problem Definition, Analytical Thinking and Decision-Making, Level 4, 2 Credits.
  • 251201-002-00-PM-03 Access, Analyse and Visualise Structured Data Using Spreadsheets, Level 4, 4 Credits.
  • 251201-002-00-PM-04 Use SQL to Communicate with a Database, Level 5, 4 Credits.
  • 251201-002-00-PM-05 Build a simple AI solution using Python, Level 5, 8 Credits.
  • 251201-002-00-PM-06 Use Python Data Scraping to Populate Database Table in SQL,
    Level 5, 4 Credits.
  • 251201-002-00-PM-07 Use Machine Learning to Build an AI solution in Python, Level 5,
    6 Credits.
  • 251201-002-00-PM-08 Use Deep Learning to Build an AI Neural Network Architecture in Python, Level 5, 10 Credits.
  • 251201-002-00-PM-09 Use Deep Learning to Build an AI Neural Network Architecture in TensorFlow, Level 5, 10 Credits.
  • 251201-002-00-PM-10 Function Ethically and Effectively as a Member of a Multidisciplinary Team, Level 4, 3 Credits.
  • 251201-002-00-PM-11 Participate in a Design Thinking for Innovation Workshop, Level 4,
    4 Credits.

    Total number of credits for Practical Skill Modules: 63

    Work Experience Modules
  • 251201-002-00-WM-01 AI Solution Design Interpretation and Development, Level 5,
    20 Credits.
  • 251201-002-00-WM-02 AI Solution Performance Testing, Level 5, 20 Credits.
  • 251201-002-00-WM-03 AI Solution Deployment, Modification and Improvement, Level 5,
    20 Credits.

    Total number of credits for Work Experience Modules: 60 
    		•

    EXIT LEVEL OUTCOMES 
    1. Gather and interpret data from various sources to define an AI solution to a real-life world problem.
    2. Critically analyse data and create a solution design document (SDD) that defines an artificial intelligence (AI) solution that solves a real-life world problem.
    3. Choose a type or category of AI learning and the relevant algorithm to analyse data, gain insight and make subsequent prediction, or create a determination with it.
    4. Train the AI model through a machine learning process and ensure that model accuracy is strictly maintained within the selection framework.
    5. Select a machine learning system and build an artificial intelligence (AI) solution to a real-life world problem.
    6. Implement and run the AI solution on a selected platform and check the prediction results in real-life use, then select and run the AI solution on a platform. 

    ASSOCIATED ASSESSMENT CRITERIA 
    Associated Assessment Criteria for Exit Level Outcome 1:
  • Verify sources from where data can be collected and identified the quantitative and qualitative quality of such data.
  • Identify a critical thinking- and problem-solving process is implemented through which the data can be reviewed for the purpose of arriving at an informed conclusion and its significance and implications.
  • Implement a critical thinking and problem-solving process through which the data can be reviewed for the purpose of arriving at an informed conclusion and identify its significance and implications.
  • Decide the scale of measurement for the data as this will have a long-term impact on data interpretation Return on Investment (ROI).
  • Prepare data by cleaning, moving, checking, and organising such data using the appropriate tools.

    Associated Assessment Criteria for Exit Level Outcome 2:
  • Write a query to extract data from an operational platform and place in a flat file (comma separated file/CSV file) or in a spreadsheet to determine what the data looks like, what type of data it is and how it ties up with other tables.
  • Use a software tool to look at and interpret data structure and the relationship between data.
  • Analyse datasets (resultant data from a query) using the appropriate tools and appropriate criteria:
  • > Access data.
  • > Add over time on new datasets.
  • > How much effort is required to clean/organise data into a usable set.
  • > Create a Software Design Document (SDD) that will describe the AI solution envisioned for the real-life world problem identified through data analysis.

    Associated Assessment Criteria for Exit Level Outcome 3:
  • Analyse data taking into consideration, the problem type is determined and classified to match the data set (groupings of data either at source or destination) to the identified problem.
  • Determine and classify the type of problem to match the data set to the identified problem, taking the data analysed into consideration.
  • Solve the type of problem, such as classification, regression, anomaly detection or determine dimensionality reduction and determine the AI algorithm that works best for each type of problem.
  • Apply basic machine learning types to determine the algorithm that should be used, i.e., supervised-, unsupervised-, semi-supervised- and reinforced learning.
  • Decide on the level of visibility needed in the AI solution and choose either decision trees or neural networks.
  • Use and compare a few different algorithms to determine which delivers the most accurate results and select an algorithm for employment.

    Associated Assessment Criteria for Exit Level Outcome 4:
  • Train the chosen algorithm through machine learning by using the data and by incrementally improving the predictions within the selection framework.
  • Maintain accuracy within the selected framework by setting minimal acceptable thresholds for the application and applying statistical discipline in the training of the algorithm to ensure accuracy.
  • Train and retrain retain AI system to ensure that the algorithm achieve the desired accuracy.
  • Test the trained models through cross-validation or by splitting the dataset (70:30) of which a portion of the data is devoted to training and the rest to testing.
  • Validate the model using the chosen metric or combination of metrics to measure the objective performance of a model.

    Associated Assessment Criteria for Exit Level Outcome 5:
  • Utilise Python's framework and libraries to solve common programming tasks and simplify the development process.
  • Determine the machine learning libraries to be used in building the AI model, taking into account that a good set of libraries means less time writing the algorithm and more time actually building the AI model.
  • Select a machine learning platform that will ease the machine learning process and facilitate building the models to build an AI system.

    Associated Assessment Criteria for Exit Level Outcome 6:
  • Plan, prepare and execute the implementation process before implementation by creating a plan to describe all tasks to be completed.
  • Minimise the ongoing risk of implementing a new solution by keeping, maintaining and monitoring a risk schedule.
  • Deploy functional resources to key locations to provide on-site support at an implementation.
  • Launch the AI solution ensuring the implementation is operating according to set outcomes and criteria.
  • Support is provided at implementation and knowledge on whether intended solution is achieved is shared.

    Integrated Assessment:
    Integrated Formative Assessment
    The skills development provider will use the curriculum to guide them on the stipulated internal assessment criteria and weighting. They will also apply the scope of practical skills and applied knowledge as stipulated by the internal assessment criteria. This formative assessment together with work experience leads to entrance in the integrated external summative assessment.

    Integrated summative assessment
    An external integrated summative assessment, conducted through the relevant QCTO Assessment Quality Partner, is required for the issuing of this qualification. The external integrated summative assessment will focus on the exit level outcomes and associated assessment criteria.

    The external integrated summative assessment will be conducted through a theoretical assessment and the evaluation of practical tasks at decentralised approved assessment sites in a simulated environment and conducted by an assessor(s) registered with the relevant AQP. 
    		•

    INTERNATIONAL COMPARABILITY 
    This qualification was compared to the following international qualifications:
    The Certificate Program in Artificial Intelligence offered by Grey Campus (with offices all over the world, including Toronto, Dubai and Mumbai) in collaboration with IBM.

    The Certificate Program in Artificial Intelligence from Grey Campus is a blend of self-paced content with live-online classes including lab and project work as practical. The program combines Data Science, Machine Learning, Deep Learning and Artificial Intelligence and focusses on a solid understanding of the tools and models. In the case of the Certificate Program in Artificial Intelligence the target group is specified as Business analysts, Information architects, qualified learners and other experienced professionals, thus indicating prior learning and experience in the IT field. The level and duration (credits) are not specified other than a reference to "self-paced". The duration is not specified other than a reference to "self-paced".
    Specific objectives covered include:
  • Understanding data structure, data manipulation and machine learning libraries.
  • Mastering machine learning concepts and techniques working with actual data.
  • Developing algorithms through supervised and unsupervised learning, performing classification and regression operations, and constructing time series models.
  • Exploring the libraries and functionalities the Python programming language offers for machine learning techniques to draw conclusions from data.
  • Understanding the concepts of supervised and unsupervised learning models together with TensorFlow, its functions, the operations and execution.
  • Deep Learning includes fundamental concepts of neural networks for building deep learning models, interpreting the result and building deep learning projects.
  • A detailed look at applying machine learning to process large natural language data.
  • Convolutional Neural networks.

    Differences:
    In the case of the Certificate Program in Artificial Intelligence the target group is specified as Business analysts, Information architects, qualified learner and other experienced professionals, thus indicating prior learning and experience in the IT field while the Occupational Certificate: Artificial Intelligence Software Developer focusses on school leavers, qualified learners from TVET colleges, new entrants into the sector and existing employees who have experience in this field, but without formal recognition of skills and competencies. Professionals who want to augment their careers may also access this qualification. Furthermore, the Certificate Program in Artificial Intelligence is self-paced and there are no prerequisites, and no prior knowledge is assumed. With the Occupational Certificate: Artificial Intelligence Software Developer there is workplace learning regulated for learners to gain meaningful practical experience under supervision and guidance.

    Similarities:
    Certificate Program in Artificial Intelligence is a foundational qualification (thus entry level). Furthermore, it states that it covers all the basics before moving to more complex aspects. Both qualifications include theory/knowledge and practical sessions. The Occupational Certificate: Artificial Intelligence Software Developer covers similar topics as the Certificate Program in Artificial Intelligence.

    United States of America:
    The second comparison was drawn with a Microsoft standard (certification exam), namely the Microsoft Certified: Azure AI Fundamentals and a preparatory qualification that is presented by Cloud Academy in San Francisco.

    Microsoft Certified: Azure AI Fundamentals leads to a certification exam according to a Microsoft specific standard. Prerequisites stated for the Azure AI Fundamentals certification include foundational knowledge of machine learning (ML), artificial intelligence (AI) concepts and related Microsoft Azure services. The target group is specified as persons with technical and non-technical backgrounds. Data science and software engineering experience are not required; however, some general programming knowledge or experience would be beneficial. The qualification is designed as a blended learning experience that combines instructor-led training with online materials on the Microsoft Learn platform and includes hands-on exercises.

    Learners are tested to demonstrate competencies in:
  • Introduction to AI.
  • Machine learning with hands-on examples on Azure platform.
  • Computer vision.
  • Natural Language Processing (NLP) workloads on Azure.
  • Conversational AI workloads on Azure.
  • Principles for responsible AI.

    Differences:
    Microsoft Certified: Azure AI Fundamentals qualification includes and is different to the Occupational Certificate: Artificial Intelligence Software Developer in terms of the following:
  • The duration being not specified.
  • Programming knowledge or experience being regarded as beneficial.
  • The qualification focuses on the ability to use Azure in the AI experience.
  • Some of the concepts covered in the qualification require a basic understanding of mathematics, such as the ability to interpret charts.

    Similarities:
    The compared Microsoft Certified: Azure AI Fundamentals and the Occupational Certificate: Artificial Intelligence Software Developer cover similar content and both include a practical aspect.

    Conclusion
    The South African Occupational Certificate: Artificial Intelligence Software Developer qualification compares favourably with the competencies covered in USA and UAE qualifications. 
    		•

    ARTICULATION OPTIONS 
    This qualification provides opportunities for horizontal and vertical articulation options.

    Horizontal Articulation:
  • Occupational Certificate: Computer Technician; NQF Level 5.

    Vertical Articulation:
  • Diploma in Information Technology; NQF Level 6. 
    		•

    NOTES 
    Qualifying for External Assessment:
    To qualify for an external assessment, learners must provide proof of completion of all required knowledge and practical modules by means of statements of results and a record of completed work experience.

    Additional Legal or Physical Entry Requirements:
  • None.

    Criteria for the accreditation of providers
    Accreditation of providers will be done against the criteria as reflected in the relevant curriculum on the QCTO website.

    The curriculum title and code are: Artificial Intelligence Software Developer: 251201-002-00-00.

    Encompassed Trade:
    This qualification encompasses the following trades as recorded on the NLRD:
  • This is not a trade qualification.

    Assessment Quality Partner (AQP)
  • MICT SETA.

    LEARNING PROGRAMMES RECORDED AGAINST THIS QUALIFICATION:
  • None. 
    		•

    LEARNING PROGRAMMES RECORDED AGAINST THIS QUALIFICATION: 
     
    NONE 


    PROVIDERS CURRENTLY ACCREDITED TO OFFER THIS QUALIFICATION: 
    This information shows the current accreditations (i.e. those not past their accreditation end dates), and is the most complete record available to SAQA as of today. Some Primary or Delegated Quality Assurance Functionaries have a lag in their recording systems for provider accreditation, in turn leading to a lag in notifying SAQA of all the providers that they have accredited to offer qualifications and unit standards, as well as any extensions to accreditation end dates. The relevant Primary or Delegated Quality Assurance Functionary should be notified if a record appears to be missing from here.
     
    NONE 


    All qualifications and part qualifications registered on the National Qualifications Framework are public property. Thus the only payment that can be made for them is for service and reproduction. It is illegal to sell this material for profit. If the material is reproduced or quoted, the South African Qualifications Authority (SAQA) should be acknowledged as the source.