Master’s Studentship: Computer Vision for Intelligent Visual Inspection

Department: Artificial Intel & Ext Reality

Cluster/Portfolio: NextGen Ent and Institutions

Overview: The CSIR has a Master's Studentship opportunity in the Emerging Digital Technologies for Fourth Industrial Revolution (EDT4IR) Research Centre within the Next Generation Enterprises and Institutions Cluster. The incumbent will be responsible for undertaking original research and development in computer vision and artificial intelligence (AI), with a specific focus on the intelligent visual inspection. The aim is to develop a modular, AI-powered platform for automated defect detection across multiple industrial domains, leveraging foundation models such as Vision Transformers (ViT) and large segmentation models. This position is based in Pretoria and will report to the Research Group Leader. Research areas: Foundation models for defect segmentation and anomaly detection (e.g. SAM, SAM2, Vision Transformers) Domain adaptation and transfer learning for industrial inspection tasks Explainability and interpretability methods for AI-based inspection systems Dataset preparation, annotation workflows, and model benchmarking for defect detection Real-time deployment of segmentation models for industrial computer vision Multimodal data integration and AI-assisted quality control systems

The Council for Scientific and Industrial Research (CSIR) is a leading scientific and technology research organisation that researches, develops, localises and diffuses technologies to accelerate socio-economic prosperity in South Africa. The organisation's work contributes to industrial development and supports a capable state.

Master's Studentship: Computer Vision for Intelligent Visual Inspection 

About the job: 

The CSIR has a Master's Studentship opportunity in the Emerging Digital Technologies for Fourth Industrial Revolution (EDT4IR) Research Centre within the Next Generation Enterprises and Institutions Cluster. The incumbent will be responsible for undertaking original research and development in computer vision and artificial intelligence (AI), with a specific focus on the intelligent visual inspection. The aim is to develop a modular, AI-powered platform for automated defect detection across multiple industrial domains, leveraging foundation models such as Vision Transformers (ViT) and large segmentation models. This position is based in Pretoria and will report to the Research Group Leader.

Research areas:

• Foundation models for defect segmentation and anomaly detection (e.g. SAM, SAM2, Vision Transformers)
• Domain adaptation and transfer learning for industrial inspection tasks
• Explainability and interpretability methods for AI-based inspection systems
• Dataset preparation, annotation workflows, and model benchmarking for defect detection
• Real-time deployment of segmentation models for industrial computer vision
• Multimodal data integration and AI-assisted quality control systems

Key responsibilities:

• Prepare research proposal to be approved by both the CSIR and the university;
• Prepare a written literature overview of the current state-of-the-art around the research topic;
• Perform original research to solve the open research problem that the student identified in his/her proposal;
• Produce sufficient quantity and quality of peer-reviewed publications around this research topic;
• Compile the master's thesis and defend the work successfully;
• Contribute to extra activities that may be outside the scope of the master's research as per the needs of the research group. The amount of such work shall be limited to 20% of the total time.

Qualifications, skills and experience:

• An Honours degree in computer science/engineering, electrical/electronic engineering, information technology or a related field;
• The ability to generate research questions and hypotheses: Review, critique and synthesize a body of research, identifying significant gaps in knowledge, methods, and study subjects to develop research questions and testable hypotheses;
• Effectively communicate issues, research findings and implications of research verbally and in writing to appropriate professional and public audiences;
• Enthusiastic and self-motivated individuals who take ownership and initiative and have the willingness to learn;
• The ability to work well under pressure as part of a team and as well as independently 
• Foundational knowledge of machine learning, deep learning, and computer vision; familiarity with Python and common ML frameworks (e.g. PyTorch, TensorFlow) will be advantageous;
• A good peer-reviewed publication record will be advantageous. 

Applicants must attach the following documents:

• An abstract of research statement/philosophy (maximum 2 pager) that highlights the problem statement, research questions, research objectives and expected outputs/outcomes/impact of the research;
• Certified copies of the applicant's latest academic results or transcript;
• Certified copies of tertiary qualifications;
• Curriculum vitae (CV);
• If applicable, proof of registration from university or confirmation letter from university supervisor that he/she is willing to be involved in the candidate supervision.

What's in it for you? 

The CSIR offers:

• Learning and development opportunities.
• Competitive and market-related remuneration packages.
• Total guaranteed salary packages which encompass contributions to the pension fund, risk insurance and medical scheme.
• We openly acknowledge and reward our employees' contributions and accomplishments through our esteemed CSIR Excellence Awards. 

PLEASE NOTE THAT FEEDBACK WILL BE GIVEN TO SHORTLISTED CANDIDATES ONLY.

For more info, please contact the CSIR Recruitment Centre on 012 841 4774 or email us at [email protected] 

The CSIR is an equal opportunity employer. As such, it is committed to the Employment Equity Act and will through the filling of this vacancy, give preference to candidates from Designated Groups in terms of the Employment Equity Act. By applying for this position at the CSIR, the applicant understands, and agrees that the CSIR may solicit a credit and criminal report from registered credit bureau and/or South African Police Services (in relation to positions that requires trust and honesty and/or entail the handling of cash or finances), and may also verify the applicant's educational qualification and employment history. The CSIR reserves the right to remove the advertisement at any time before the stated closing date and it further reserves the right not to appoint if a suitable candidate is not identified.