SDA-4 Applied AI Integration
SDA-4 Applied AI Integration
- Course description
- Course CodeSDA-4
- Level of Study5.1
- Program of StudySoftware Development with AI
- Credits10
- Study Plan CoordinatorMonde Sineke
This course introduces students to integrating AI capabilities into existing software applications. Students learn how language models work at a practical level, including their capabilities, limitations, cost and non-deterministic behaviour, and how to select a model that fits an application's requirements. The course covers prompt design, context management, structured outputs and the implementation of AI-powered application features, including error handling and validation.
Through practical development activities, students integrate an AI service to address a need and evaluate its outputs for accuracy, reliability and usefulness. The course introduces common risks associated with AI integration, including hallucinations, prompt injection and data protection concerns, alongside principles of transparency, human oversight and responsible AI use. Because AI models and services change quickly, students also learn how to follow developments and reassess earlier choices.
The candidate...
- has knowledge of how AI models work at a practical level, including their capabilities, limitations, cost and non-deterministic behaviour
- has insight into data protection and security requirements when integrating AI services, including the implementation of appropriate guardrails to mitigate risks such as prompt injection.
- can update their knowledge of AI models and services, and reassess earlier choices as they change
The candidate...
- can apply knowledge of AI models to select an appropriate model for an application based on its requirements, performance and cost
- can apply prompt design, context management and structured output to integrate an AI service into an existing application, with appropriate error handling
- can assess the output of an AI feature for accuracy, reliability and usefulness, identify failures such as hallucinations, and determine what measures need to be implemented
The candidate...
- understands the ethical principles that apply to AI features in software, including transparency, data protection and human oversight
Students use a combination of written content, videos and practical learning activities designed to help them achieve the course learning outcomes.
Industry Relevance
Course content, learning activities and tools reflect industry standards and are developed in collaboration with industry partners and experts from the field.
Practical Work
The course includes regular practical activities where students apply their knowledge and practise relevant skills.
Collaboration
Collaboration is facilitated through the LMS and Teams, and students are encouraged to work together for peer support and shared learning.
Guidance
Teachers are available for guidance at scheduled times. Students are encouraged to share work in progress and use feedback from teachers and fellow students to develop their work further.
| Form of assessment | Grading scale | Grouping | Duration of assessment |
|---|---|---|---|
Course Assignment | Pass / Fail | Individual | 1 Week(s) |