Higher Diploma in AI Systems Engineering
Every organisation in the region is now buying AI, and almost none can staff it: banks, telecoms, government programmes, health systems and the donor funded sector all procure AI systems they cannot specify, evaluate, deploy or govern.
Occupational profile
What you'll be paid to do
Every organisation in the region is now buying AI, and almost none can staff it: banks, telecoms, government programmes, health systems and the donor funded sector all procure AI systems they cannot specify, evaluate, deploy or govern.
The result is expensive imported systems that fail in local conditions and sit unused.
The missing role is the AI systems engineer: the person who builds, deploys, operates and governs these systems, and who can tell a client honestly when the answer is not to build one.
The graduate works as any of the following, in the region's highest paying technical roles: engineer technology companies, monthly; remote rates Data engineer Any organisation with data UGX 2,000,000 to 4,500,000 technical lead startup economy the highest ceiling in the
| Occupation | Where the work is | Starting earnings, observed 2026 |
|---|---|---|
| AI or machine learning engineer | Banks, telecoms, technology companies, remote international clients | UGX 2,500,000 to 6,000,000 monthly; remote rates substantially higher |
| Data engineer | Any organisation with data at scale; consultancies; remote clients | UGX 2,000,000 to 4,500,000 monthly |
| AI solutions consultant | Consultancies, donor programmes, government digital units | UGX 2,000,000 to 5,000,000 monthly; project rates |
| Technology founder and | Own venture; the regional | Equity and venture returns; |
The 90-Day Promise
From certification to income, stated concretely
Day Pathway Named employer categories with which the College holds or is concluding placement and hiring partnerships: banks and fintechs; telecoms; technology organisations; donor programmes; and regional and international employers offering remote engagement.
Partnership target: twenty active partners by the first cohort's year two training, each offering at least two placements, with at least five offering remote work.
The ninety day pathway from certification to income:
Day 0: graduation with the Capability Transcript, the mentor's reference, a deployed capstone system with published evaluation, a public repository portfolio and the external panel's assessment.
Day 0: graduation with the Capability Transcript, the mentor's reference, a deployed capstone system with published evaluation, a public repository portfolio and the external panel's assessment.
Days 1 to 30: employed route: three arranged interviews in the region's tightest technical labour market. Remote and export route: profiles live on international contracting platforms with the capstone as portfolio evidence; the mentor reviews rate setting.
Days 31 to 90: employed route: probation support from the practice tutor. Venture route: the venture's users and traction continue; referral listing for partner overflow engineering work, standing terms in the partnerships.
Measurement: employment, remote contracting or trading status recorded at day 90 and published in the annual outcomes report, per Part Two, section 2.9.
Programme structure
Every module, with its arithmetic
One credit unit equals ten notional learning hours. 240 credit units, 2400 notional hours: 685 contact, 1380 practical and 330 industrial hours, so practical and industrial hours are 71 per cent of the programme.
h and and ce and nt One y and ial Two Practical and industrial hours together are 1710 of 2400 notional hours, 71 per cent, meeting the hands on test.
Semester 1
| Code | Module | CU | Contact h | Practical h | Industry h |
|---|---|---|---|---|---|
| AIS301 | Machine Learning Systems in PracticeAbove the diploma's tool use sits the engineer who understands what the model is doing and why it fails: the learner builds, trains and evaluates machine learning systems on real data with honest measurement. | 14 | 50 | 90 | 0 |
| AIS302 | Data Engineering and PipelinesMost AI failures are data failures: the learner builds the pipelines that move, clean and serve data reliably, the unglamorous work that decides whether any system works at all. | 14 | 45 | 95 | 0 |
| AIS303 | Software Engineering for AI SystemsAI in production is software: the learner writes, tests and structures code to the standard that lets systems be maintained by someone other than their author. | 14 | 50 | 90 | 0 |
| RES301 | Applied Research and Evidence PracticeThe Level 6 difference is evidence: the technologist who can frame a question, gather data honestly and defend a conclusion is the one trusted with decisions. This module builds that discipline in applied, workplace terms. | 10 | 45 | 55 | 0 |
Semester 2
| Code | Module | CU | Contact h | Practical h | Industry h |
|---|---|---|---|---|---|
| AIS304 | Deploying and Operating AI SystemsA model in a notebook earns nothing: the learner deploys AI systems into production and keeps them running, monitored and improving, which is where most of the engineering actually lives. | 14 | 45 | 95 | 0 |
| AIS305 | Applied Natural Language and Vision SystemsThe two application areas that carry most commercial value: the learner builds language and vision systems for real problems, including the African language and local imagery contexts that global systems serve poorly. | 14 | 45 | 95 | 0 |
| AIS306 | AI Safety, Ethics and GovernanceThe engineer who cannot reason about harm should not be trusted to build: this module makes safety, fairness and governance an engineering competence rather than a compliance afterthought. | 14 | 60 | 80 | 0 |
| SPNH101AI spine | Frontier AI Practice and Systems JudgementThe Higher Diploma spine: the graduate must stay current in a field that changes yearly and must exercise the judgement about when and whether to build that defines a senior engineer. | 16 | 55 | 105 | 0 |
| ITRH101 | Industrial Training OneTen assessed weeks in an engineering or technology organisation at the end of year one, at Level 6 responsibility, under the Part Two policy. | 16 | 0 | 10 | 150 |
Semester 3
| Code | Module | CU | Contact h | Practical h | Industry h |
|---|---|---|---|---|---|
| AIS401 | AI Systems Architecture and Scale | 18 | 55 | 125 | 0 |
| AIS402 | AI SolutionsYear two's core: the learner designs and builds AI systems that serve real load, integrate with organisational infrastructure and can be operated by a team, which is the Level 6 engineering task. | 16 | 50 | 110 | 0 |
| AIS403 | Advanced Practice ElectiveDepth beats breadth at Level 6: the learner specialises in one strand to genuine professional depth, building the | 16 | 50 | 110 | 0 |
| LEAD301 | Technical Leadership and Project ManagementThe Higher Diploma exists to produce the person who runs the team and the project: this module builds the leadership, planning and financial control that separates a technologist from a technician. | 12 | 50 | 70 | 0 |
Semester 4
| Code | Module | CU | Contact h | Practical h | Industry h |
|---|---|---|---|---|---|
| SPN102AI spine | Digital Work and Platform IncomeThe shortest module with the fastest payback: how a technician finds customers, prices work, invoices and gets paid through digital channels. | 4 | 15 | 20 | 0 |
| ENT301 | Technology Enterprise and Commercial PracticeAt Level 6 the enterprise question changes from can you trade to can you build something that scales: this module treats the technology business, its finance, contracts and growth, at the depth the graduate's ventures now require. | 10 | 40 | 60 | 0 |
| ITRH102 | Industrial Training TwoTwelve assessed weeks in year two at technologist level with a substantial host project, under the Part Two policy. | 20 | 0 | 20 | 180 |
| EXPH101 | Capstone Engineering ProjectThe Level 6 closing argument: a complete AI system engineered, deployed, evaluated and defended, at a scale and rigour that a hiring engineering manager recognises as professional work. | 18 | 30 | 150 | 0 |
The AI and Digital Practice Spine
AI, taught inside this trade
Applied Here The spine of Part Two, section 2.13 at Higher Diploma depth: 20 credit units across SPNH101 and SPN102.
AI tools for my trade
The engineer builds the tools; the spine adds rigorous evaluation of every new
Learning outcomes
What the graduate can do
- PLO 1
Build, train and honestly evaluate machine learning systems on real data
Level 6: applies advanced knowledge with critical evaluation
- PLO 2
Design and operate data pipelines with quality monitoring and lawful data handling
Level 6: manages complex technical systems
- PLO 3
Engineer software to production standard with testing, review and architecture discipline
Level 6: applies specialised technical expertise
- PLO 4
Deploy, monitor, operate and improve AI systems in production with cost control
Level 6: manages systems with accountability for outcomes
- PLO 5
Build applied natural language and computer vision systems including for local contexts
Level 6: applies advanced skills to complex problems
- PLO 6
Assess AI harms, test for fairness and robustness, and govern systems under regulation
Level 6: exercises professional judgement on complex ethical matters
- PLO 7
Architect AI systems for scale, integration and operability by a team
Level 6: designs complex systems
- PLO 8
Engineer AI solutions for African constraints and evaluate their impact honestly
Level 6: applies expertise to contextual problems with critical evaluation
- PLO 9
Evaluate emerging AI capabilities rigorously and exercise build or decline judgement
Level 6: critically evaluates new knowledge
- PLO 10
Conduct applied research and defend evidence based conclusions
Level 6: investigates and reports with rigour
- PLO 11
Lead technical teams and manage projects to scope, budget and schedule
Level 6: manages resources and people with accountability
- PLO 12
Build and operate a technology enterprise with defensible economics, contracts and growth planning
Level 6: operates autonomously with commercial responsibility
Entry routes
Four ways in
- Academic route. A Diploma at UVQF Level 5 in computing, engineering, mathematics or a cognate field, or an equivalent qualification.
- Vocational route. FC-D01 Applied AI and Data Operations is the designed feeder, with its 240 credit units recognised toward cognate modules; FC- D06 and FC-D11 graduates enter with bridging in AIS301.
- Recognition of prior learning. Practising software and data professionals assess against year one outcomes; advanced standing follows demonstrated competence and a portfolio review.
- Mature age route. Applicants of twenty five and above with substantial technical practice and a portfolio enter through the access assessment and technical interview.
Industrial training
Assessed weeks inside a working organisation
- Duration and placement
- Ten weeks in year one (ITRH101, 16 credit units, 150 hours minimum) and twelve weeks in year two (ITRH102, 20 credit units, 180 hours minimum).
- Host organisation types
- Banks and fintechs, telecoms, technology companies and startups, government digital and data units, research organisations, donor programmes with data operations, and remote placements with regional and international employers.
- Learning objectives
- Year one: machine learning and data engineering work at technologist level under supervision. Year two: system design, deployment and governance work with a substantial host project of measured value.
- Supervision
- A named host supervisor engineer; the practice tutor visits at least twice per placement or reviews remotely for remote placements; data and governance incidents escalate immediately.
- Logbook
- Daily entries against the objectives, countersigned weekly by the host supervisor, reviewed by the practice tutor at each visit, kept on the platform.
- Assessment
- Joint workplace assessment against the Level 6 practice rubric each year, plus logbooks and the year two host project.
Assessment and certification
Continuous practical, plus UVTAB
A human assessor confirms every AI-avatar oral examination result before it stands, and the trainer who taught a learner never marks that learner's summative assessment.
- Internal continuous assessment, 60 per cent: workshop task assessments, module projects, logbooks and oral checkpoints, marked under the separation rule of Part Two.
- UVTAB external assessment, 40 per cent: the Board's written and practical occupational assessments for this qualification, taken at the gazetted sittings.
- Practical competency assessment: every practical outcome is assessed by observed performance against published criteria; evidence is retained.
- Oral examination option: any module checkpoint and the exit project defence may be taken in the AI avatar oral examination room, with human confirmation of every result.
- Grading scale: Distinction 80 to 100; Credit 65 to 79; Pass 50 to 64; Not Yet Competent below 50, with the right to reassessment.
- Pass and progression: every core module at Pass or above, all practical competencies demonstrated, industrial training completed, exit project at Pass or above.
- Resit rules: two reassessment opportunities per module without repeating attendance; a third attempt repeats the module; reassessment covers only the outcomes not yet demonstrated.
Exit project
A real venture, or a real employer brief
The capstone is a complete AI system engineered end to end for a real client or a documented real need: problem framing and the honest assessment of whether AI is the right answer, data pipeline, model or system build, production deployment, monitoring, a fairness and harm assessment with published results, governance documentation, and an impact evaluation.
Venture route candidates may instead deliver a technology venture with a deployed product, real users and documented commercial traction, to the same technical standard.
Deliverables:
The engineering file: problem framing, architecture, design decisions and the reasoning documented The deployed system: running in production, monitored, with runbooks and a handover pack The evaluation: technical performance, fairness and harm assessment, and honest impact results The defence: presentation, technical examination and oral defence before a professional panel, human confirmed Grading criterion Weight What excellent looks like Defence 15% Every technical decision
- The engineering file: problem framing, architecture, design decisions and the reasoning documented
- The deployed system: running in production, monitored, with runbooks and a handover pack
- The evaluation: technical performance, fairness and harm assessment, and honest impact results
- The defence: presentation, technical examination and oral defence before a professional panel, human confirmed Grading criterion Weight What excellent looks like Defence 15% Every technical decision
Progression
Where this qualification leads next
No Forward College learner ever meets a dead end.
Into Forward University:
FC-H01 graduates enter the University's computing, data science and artificial intelligence degrees with credit transfer of up to one half at Level 6 under the Part Two, section 2.12 equivalence tables, the most generous articulation in the compendium and the reason the award is engineered to degree adjacent rigour.
Professionally: the award is designed to be recognised directly by employers for technologist and engineer roles without further study, and the capstone panel includes external industry assessors for that reason.