Turn raw and messy data into reliable evidence, build and deploy machine learning systems responsibly, and support better decisions at home.
The Department of Data Science and Artificial Intelligence is the academic unit of the Faculty of Computing at Ilays National University responsible for teaching, research, and community engagement in the analysis of data and the design of intelligent systems. It prepares graduates who can turn raw and messy data into reliable evidence, build and deploy machine learning systems responsibly, and support decision making in organizations that are collecting far more data than they currently know how to use.
The department exists to close a specific and measurable gap. Institutions across the country generate substantial operational data through mobile money platforms, telecommunications networks, health information systems, humanitarian registries, customs and revenue systems, and routine public administration. Very little of that data is modelled locally. Where analysis is done at all, it is frequently outsourced to consultants outside the region — expensive, slow, and leaving no capability behind. The department was established to build that capability at home.
The department currently offers the Bachelor of Data Science and Artificial Intelligence — a foundation year followed by four years of degree study, comprising ten semesters and 155 credit units. The programme is delivered in close cooperation with the Bachelor of Software Engineering, with which it shares roughly three quarters of its coursework. This shared model gives students a genuine computing foundation rather than a narrow analytics training, and allows the faculty to run both degrees efficiently with a single teaching establishment.
To deliver rigorous, practice-oriented education in statistics, computing, and machine learning, taught by staff who have applied these methods to real problems.
To develop applied research that addresses problems of local consequence, including low resource language technology, health analytics, climate and agricultural forecasting, and financial integrity.
To promote the ethical, lawful, and socially responsible use of data and automated decision systems in every course the department teaches.
To build durable partnerships with industry, government, and the humanitarian sector so that student projects, internships, and research answer questions those partners actually have.
To serve as a national resource for data literacy through short professional courses, public workshops, and open teaching materials.
Six sectors have already crossed the threshold where data volume exceeds what manual reporting can handle — and each lacks the staff to act on it.
Call detail records, transaction logs, subscriber and agent networks
Core banking records, remittance flows, loan performance
Facility reporting, immunization and nutrition registries, disease surveillance
Rainfall and satellite records, market prices, livestock movement
Beneficiary registries, needs assessments, logistics and distribution records
Census and survey data, revenue and customs records, education statistics
The common obstacle in all six sectors is not the absence of data and not the absence of software — it is the absence of people who can frame a question statistically, prepare the data honestly, choose an appropriate method, and communicate the result to a decision maker who is not a technical specialist.
Short programmes of eight to twelve weeks in applied data analytics, business intelligence, and applied machine learning, aimed at working professionals in banking, telecommunications, and the humanitarian sector.
A conversion route for graduates of other computing and quantitative disciplines who need structured postgraduate preparation in data science and AI.
A Master of Science in Data Science and Artificial Intelligence, to be introduced once the department has a sufficient body of research-active staff and senate approval.
The department also provides service teaching to other academic units, principally in probability and statistics, data analysis, and research methods, and supports quantitative components of programmes in business, health sciences, and social sciences.
What graduates are expected to have achieved three to five years after completing the degree.
| Ref. | Objective | Statement |
|---|---|---|
| PEO 1 | Professional Practice | Graduates practice competently as data scientists, data analysts, data engineers, or machine learning engineers in industry, government, or the development sector, and take responsibility for the technical quality of their work. |
| PEO 2 | End-to-End Solutions | Graduates design and deliver end to end data and AI solutions, from problem framing and data acquisition through modelling to deployment, monitoring, and communication of results. |
| PEO 3 | Ethical Judgement | Graduates exercise professional and ethical judgement, recognizing the legal, privacy, fairness, and social implications of the systems they build. |
| PEO 4 | Lifelong Learning | Graduates continue to learn independently, adopting new methods and tools as the field changes, and a proportion progress to postgraduate study or research. |
| PEO 5 | Capability Building | Graduates contribute to building data capability in the institutions they join, mentoring colleagues and improving how those institutions use evidence. |
On successful completion of the degree, graduates will be able to:
Apply calculus, linear algebra, probability, and statistical theory to formulate and solve data driven problems.
Write correct, efficient, and maintainable programs in C, Java, and Python, and select appropriate data structures and algorithms for a given computational task.
Design, query, and manage relational and non-relational data stores, and build data pipelines that acquire, clean, transform, and integrate data from heterogeneous sources.
Conduct exploratory and confirmatory statistical analysis, including estimation, hypothesis testing, regression, and the design of experiments and controlled trials.
Select, train, tune, and rigorously evaluate supervised and unsupervised machine learning models, and explain the assumptions and failure modes of each method used.
Design and train deep neural network architectures and apply them to problems in natural language, vision, and sequential data.
Build, analyse, and forecast time dependent data using classical and machine learning approaches.
Deploy models into production environments and monitor them for performance degradation, data drift, and operational failure.
Communicate quantitative findings clearly through visualisation, written reports, and oral presentation to both technical and non-technical audiences.
Evaluate the ethical, legal, privacy, and fairness implications of data collection and automated decision making, and apply appropriate safeguards.
Plan and conduct an independent investigation, including a literature review, a defensible methodology, and a critical discussion of results and limitations.
Work effectively in a team, manage a project against a schedule, and apply professional software engineering practice to data and AI systems.
Ten semesters, 155 credit units. Progression is deliberately sequenced — no specialisation course is taught before its mathematical prerequisite has been examined.
| Year | Semester I | Semester II | Year Total |
|---|---|---|---|
| Foundation Year | 16 | 20 | 36 |
| Year I | 18 | 15 | 33 |
| Year II | 15 | 14 | 29 |
| Year III | 15 | 12 | 27 |
| Year IV | 12 | 18 | 30 |
| Programme Total | 155 |
Practical hours exceed lecture hours across the pathway, and dominate in all programming, machine learning and data engineering courses.
Academic English, pre-calculus and calculus, computer fundamentals and applications, introductory programming, web and hardware basics, general education.
Programming in C and Python, discrete mathematics, probability and statistics, linear algebra, data structures and algorithms, object-oriented programming, databases, computer organisation, operating systems.
Data analysis, statistical inference and experimental design, database and data management, software engineering, systems analysis, algorithms, cloud and distributed computing, machine learning foundations, ethics and entrepreneurship.
Deep learning, data engineering and big data, data visualisation, natural language processing, time series forecasting, computer vision, reinforcement learning, MLOps, distributed systems, research methods.
Generative AI and large language models, AI ethics and governance, internship, capstone project, capstone seminar and portfolio.
Twelve taught courses and three programme projects establish the distinct data science and AI identity of the degree.
Estimation, confidence intervals, hypothesis testing, ANOVA, regression diagnostics, sampling and the design of experiments and A/B tests. Builds directly on PS103.
Supervised and unsupervised learning: regression, classification, decision trees, ensembles, clustering and dimensionality reduction, with cross-validation, regularization and model evaluation in scikit-learn.
Feedforward, convolutional and recurrent architectures, backpropagation, optimizers, regularization and transfer learning, implemented in PyTorch or TensorFlow.
Data pipelines, ETL and ELT, batch and streaming processing, distributed storage, warehouses and lakes, Spark and workflow orchestration.
Principles of visual encoding, exploratory and explanatory graphics, dashboard design and communicating quantitative findings to non-technical audiences.
Text preprocessing, representation learning, sequence models, transformers, and applications such as classification, named entity recognition and question answering.
Stationarity, decomposition, ARIMA and exponential smoothing, state space models and modern machine learning approaches to forecasting.
Image formation and processing, feature extraction, convolutional networks for classification, detection and segmentation, and vision model evaluation.
Markov decision processes, dynamic programming, Monte Carlo and temporal difference methods, policy gradients and deep reinforcement learning agents.
Packaging, serving and monitoring models in production: experiment tracking, CI/CD for ML, containerization, drift detection and model lifecycle management.
Transformer based generative models, pretraining and fine-tuning, prompting, retrieval augmented generation, evaluation and the limits of generative systems.
Fairness, bias, transparency and accountability in data and AI systems, privacy preserving techniques, data protection regulation and emerging AI governance frameworks.
Teaching combines structured lectures for theory and derivation, supervised laboratory sessions in which students implement methods themselves, problem sets that require working from raw data rather than prepared extracts, case studies drawn from partner organizations, seminar discussion for the ethics and governance courses, and individual and team projects at four points in the programme. Assessment combines continuous assessment — laboratory work, assignments, quizzes, project deliverables and presentations — with end-of-semester examination, weighting continuous assessment more heavily where the practical component is large.
An individual data analytics project working from raw data through preparation, analysis and reporting.
A team machine learning project covering modelling, evaluation and collaborative delivery.
An integrated data science project spanning acquisition, modelling and deployment.
A substantial supervised piece of work, wherever possible on a real problem supplied by an external partner, defended publicly before a panel.
| Facility | Purpose | Minimum Provision |
|---|---|---|
| General computing laboratory | Foundation and Year I programming, databases, operating systems | Shared with the Faculty of Computing, one workstation per student per session |
| Data science laboratory | Data analysis, statistics, visualisation, machine learning | Dedicated laboratory, 16 GB RAM and multi-core CPU per workstation |
| AI and deep learning facility | Deep learning, computer vision, NLP, generative models | GPU server over the campus network, supplemented by cloud credits |
| Data engineering environment | Big data, distributed processing, pipelines, MLOps | Virtualised cluster supporting Spark, orchestration and container registries |
| Project and research space | Team projects, capstone work, research students | Bookable room with presentation facilities and secure storage for project data |
Deliberately open source, so graduates can reproduce their working environment anywhere and employers face no licensing barrier.
The degree is built backwards from the roles graduates are expected to enter. Specialisation courses allow students to weight their capstone toward a particular destination.
| Role | What the graduate does | Typical Employers |
|---|---|---|
| Data analyst | Prepares, analyses, and reports on operational data; builds dashboards and recurring reporting | Banks, telecommunications operators, NGOs, ministries |
| Data scientist | Frames problems statistically, builds predictive and inferential models, communicates findings | Financial services, telecommunications, research institutes |
| Machine learning engineer | Builds, trains, and optimises models and integrates them into production systems | Technology firms, fintech, larger operators |
| Data engineer | Designs and operates data pipelines, warehouses, and processing infrastructure | Any organisation with substantial data volume |
| MLOps engineer | Deploys, monitors, and maintains models in production | Technology and financial services firms |
| Business intelligence analyst | Translates business questions into measures, reports, and decision support | Commercial enterprises, public agencies |
| Monitoring & evaluation analyst | Designs indicators, analyses programme data, evaluates impact | Humanitarian and development organisations, UN agencies |
| Statistician or survey analyst | Designs samples and surveys, produces and quality assures official statistics | National statistics offices, research organisations |
| Database administrator | Designs, secures, and maintains organisational data stores | Banks, hospitals, government, telecommunications |
| Postgraduate researcher | Continues to master's or doctoral study in data science, AI, or a quantitative discipline | Universities in the region and abroad |
Corpus building, tokenization, machine translation, speech recognition and text classification for Somali and other regional languages.
Disease surveillance and outbreak forecasting, service coverage analysis and decision support for facilities operating with incomplete records.
Drought and rainfall forecasting, remote sensing for rangeland and crop monitoring, livestock market price prediction and early warning.
Credit scoring for populations without formal credit histories, fraud detection in mobile money and remittance flow analysis.
Survey methodology, record linkage, official statistics quality and open data infrastructure for public institutions.
Ethics is addressed at two points and reinforced throughout: a Year II course in computing ethics and professional practice, and a Year IV course in AI ethics, governance and data privacy. Beyond the taught courses, all student projects involving personal or sensitive data require ethical clearance before data collection or analysis begins, and project rubrics assess how data was obtained, how it was protected, who the model affects, and what the consequences of error would be.
The department maintains formal relationships with employers in telecommunications, banking and mobile money, health, humanitarian and public sectors. These supply internship places, real problems and data for student projects, adjunct staff and guest speakers, and curriculum review through the advisory board. Every student completes a supervised internship in Year IV, and the capstone seminar assembles a documented professional portfolio.
Every course maintains a file containing the syllabus, teaching materials, assessment instruments, samples of marked work across the grade range, and the lecturer\u2019s end-of-semester reflection.
Feedback is collected each semester and reviewed by the Head of Department with the course lecturer.
An advisory board drawn from industry, government and the development sector meets at least annually to review curriculum relevance.
A full curriculum review is conducted every three to four years, reflecting a field in which methods and tools change quickly.
Graduate destinations are tracked at six months and at three years, and reported against the programme educational objectives.
External examiners or peer reviewers are invited to moderate final-year assessment and capstone defences.
All student projects involving personal or sensitive data require ethical clearance before data collection or analysis begins.
Contact Admissions