Autonomously and independently evaluate deficiencies when interacting with a range of technologies and leveraging knowledge of these deficiencies to improve future practice.
COM774 (79166) Intelligence Engineering and Infrastructure
Brief: Coursework II – Set Exercise [75%]
Submission Deadline: 8th December 2025
Feedback Date: 13th January 2026
Module Title: Intelligence Engineering and Infrastructure
Module Code: COM774 (79166)
Semester (s) Taught: One Course / Year Group: MSc CS/AI/AI(CUQ)/7
Coursework / Exam Weighting: 75% (of Total coursework)
Coursework Assessment
When submitting your assignment you are agreeing to the following statement:
“I declare that this is all my own work. Any material I have referred to has been accurately referenced and any contribution of Artificial Intelligence technology has been fully acknowledged. I understand the importance of academic integrity and have read and understood the University’s General Regulation: Student Academic Integrity and the Academic Misconduct Procedure. I understand that I must not upload my work before, during or after submission to any unapproved plagiarism detectors or answer sharing platforms, or equivalent, and that only University-approved platforms should be used. A mark of zero may be awarded and the reason for that mark will be recorded on my file.”
Policies, procedures, resources etc. can be found on the Academic Misconduct SharePoint site: https://ulster.sharepoint.com.mcas.ms/sites/AcademicIntegrity.
This module is assessed by two pieces i.e. coursework CW1 and CW2.
Coursework II is explained in the document as follows:
In Coursework CW2, the focus will shift to the remaining stages of the MLOps workflow. This includes Model Development, where machine learning models are designed, trained, and validated; CI/CD, which automates testing and integration of code, data, and models; and Deployment, where models are moved into production environments. It also covers Monitoring, to track performance and detect issues such as data drift; Retraining, to update models with new data; and Governance, which ensures compliance, transparency, and accountability across the entire lifecycle.
Related Learning Outcomes:
1. Autonomously and independently evaluate deficiencies when interacting with a range of technologies and leveraging knowledge of these deficiencies to improve future practice.
2. Appraise, select and autonomously apply skills to leverage a range of machine learning paradigms.
3. Demonstrate the ability to critically appraise meant by Intelligence engineering and infrastructure and how a variety of processes and paradigms may be applied to address the challenges it presents.
Students will be set an individual exercise where they will be expected to utilize the previously identified dataset in CW1 which can be further used to produce a machine learning model to address a problem. Students will then implement a solution to the problem using technologies and techniques covered in the module.
Requirements:
Specifically for this exercise, students are required to perform the following 3 tasks.
Task 1: Solution Design
Using the dataset identified in CW1 (e.g., human activity recognition dataset), design a solution that applies MLOps processes and technologies. Your design should include:
- Use of the pre-selected dataset from CW1.
- Appropriate data versioning and storage platforms.
- Training a machine learning model and evaluating its performance across at least two iterations of the pipeline, using tools for managing and tracking experiments.
- Evaluation/testing methods to detect performance regressions.
- Consideration of scalability and deployment constraints.
Task 2: Implementation, Deployment, and Testing
- Implement and deploy the designed solution using the technologies taught in the module.
- Consider limitations of free cloud tiers and potential costs.
- Important: If using cloud services, you must use the platform covered in the module’s teaching materials.
Task 3: Presentation & Demonstration
- Slide 0: Title slide (project name, one-line description, student name, student number).
- Slides 1–2: Problem discussion and relevant MLOps considerations.
- Slides 3–6: Solution design and development process.
- Slides 7–8: Key features (e.g., model testing, evaluation methods).
- Slides 9–10: Limitations and scalability evaluation.
- Slide 11: Recorded demonstration of solution functionality.
- Slide 12: Provide concluding comments and a properly formatted reference list.
- Clearly demonstrate the solution.
- Showcase the use of MLOps tools, testing, and deployment.
- P]4]rovide evidence of decisions and processes referenced in the slides.
Assessment Criteria: Your submission will be assessed against the following criteria:
1. Problem Definition and Discussion (10%)
- Quality of the description of the overall problem.
- Justification of why scalable intelligence engineering is needed.
- Critical appraisal of relevant MLOps and intelligence engineering patterns.
2. Overview of the Technical Solution Developed (15%)
- Justification for chosen technologies and design.
- Consideration of alternative technologies.
- Incorporation of scalable elements and design patterns.
- Quality of solution architecture and diagrams.
3. Testing Approaches (20%)
- Integration of testing into the solution.
- Use of appropriate metrics.
- Application of testing to model evaluation and pipeline refinement.
4. Performance Evaluation and Scalability (20%)
- Assessment of solution limitations.
- Strategies to address weaknesses.
- Evaluation of scalability across deployment environments.
5. Concluding Comments, References & Presentation (10%)
- Quality of reflection on functionality, limitations, and applicability.
- Accuracy and formatting of references.
- Professionalism of presentation (structure, clarity, formatting, figures).
6. Video Demonstration (25%)
- Functionality of the solution.
- Clarity and quality of demonstration.
- Evidence of using scalable intelligence engineering infrastructure.
Note: Total: 100% (equivalent to 75% of module mark)
Submission Guidelines:
Prepare your presentation slides.
- Create your presentation slides in PowerPoint format (.pptx).
- Ensure slides are clear, well-structured, and support the flow of your video presentation.
Record your video presentation.
- Record a 5-minute video presentation incorporating your PowerPoint slides.
- The video should clearly explain the complete MLOps workflow, including the data pipeline (covered in CW1), model development, continuous integration and deployment (CI/CD), deployment, monitoring, retraining, and governance.
- Ensure both audio narration and visuals (slides, code examples, or demonstrations) are clear and professional.
Combine slides and video
- Save your presentation as a single file that includes both your video recording and slides.
- Check playback to ensure that audio, visuals, and timing work correctly.
Final checks
- Verify that your submission is complete, professional, and within the time limit.
- Ensure your name, student ID, and module code are included on the first slide.
- Confirm that the file format is accepted (PowerPoint with embedded video).
Upload to Blackboard
- Log in to Blackboard and navigate to the Assessment 1 submission folder. Upload your final presentation file before the submission deadline.
- Double-check that the correct version has been uploaded.
Plagiarism and academic integrity
- Ensure your submission complies with the University’s plagiarism policy.
- Any use of external sources, datasets, or code must be appropriately referenced.
Note: According to Ulster University Assessment Code of Practice, where submitted work exceeds the agreed assessment limit, a margin of up to +10% of the work limit will be allowed without any penalty of mark deduction. If the work submitted is significantly in excess of the specified limit (+10%), there is no expectation that staff will assess the piece beyond the limit or provide feedback on work beyond this point. Markers will indicate the point at which the limit is reached and where they have stopped marking. A mark will be awarded only for the content submitted up to this point. No additional deduction or penalty will be applied to the overall mark awarded. The student is self-penalising as work will not be considered/marked.
N.B. Students should be aware of the plagiarism policy of the University and submit their coursework in accordance with this.
N.B. The students are required to implement this solution using the concepts and techniques which were the focus of the teaching materials in this module. This may broadly have a focus on a hosted/cloud native design or a containerised solution. It is recommended that students appraise both these design approaches in their slides.
