ICT Career Guide · Kenya · Article 201
12 ICT Portfolio Projects for Jobs in Kenya
Choose projects that prove what your target ICT job actually asks for—then test and present the evidence clearly.
Which ICT portfolio projects can help you get hired in Kenya?
Choose projects that match the work you want. Strong options include an inventory system, booking system, order tracker, Kenya public-data analysis, business sales analysis, cybersecurity monitoring lab, secure-application assessment, AI document assistant, prediction project, mobile expense tracker, offline check-in app and one job-targeted flagship system.
You do not need all twelve. GitHub recommends highlighting about three to five relevant projects and making them easy for hiring managers to understand quickly. [1]
Choose fewer projects, finish them properly
Start with a three-project portfolio strategy
If you are a beginner, a focused portfolio is easier to finish and easier for a reviewer to understand. Build three projects with different jobs.
Foundation project
Prove that you understand the basics: a responsive site, small database app, simple analysis, mobile utility or controlled cyber lab.
Role-specific project
Choose a project that closely matches the job family you want—for example SQL and dashboards for data, or APIs and authentication for development.
Flagship project
Build one complete workflow that you can demonstrate, test and defend in an interview. Add more projects only when they prove something meaningfully different.
A project earns space in your portfolio when it proves a capability that matters to the role you want—not merely because its title sounds advanced.
Start with evidence
Match the project to the job you want
GitHub recommends highlighting a small set of relevant projects for hiring review and making each project easy to understand through clear repository documentation, demonstrations and evidence of testing. [1] [2]
| If the job repeatedly asks for… | Build evidence such as… | A reviewer should be able to inspect… |
|---|---|---|
| SQL, dashboards, data quality | End-to-end analysis using raw data, cleaning, SQL/Python/Excel and a dashboard | Queries, validation, calculations, dashboard and conclusions |
| APIs, databases, authentication | Complete web application with persistent data and permissions | Endpoints, schema, login/roles, error handling and tests |
| Mobile development | Working app with storage, validation and offline behaviour | App flow, persistence, edge cases and device demonstration |
| AI / LLM applications | AI system with retrieval or model logic plus evaluation | Inputs, source data, expected output, failures and improvements |
| Cybersecurity monitoring | Authorised lab with logs, detection and investigation | Event, evidence, query/rule, finding and response |
| Git / GitHub | Well-structured repository and documentation | README, meaningful history, setup steps and tests |
On a phone, swipe horizontally to view all three columns. For broader employability skills rather than project selection, use NIT’s ICT Skills Employers Want in Kenya guide. Article 201 stays focused on project evidence.
Before choosing a project
What makes an ICT project worth putting in your portfolio?
It solves something you can explain
State the problem in one sentence. Start with the user’s need rather than a list of technologies.
It has a clear person or organisation using it
Knowing the user makes requirements, permissions, inputs and outputs easier to define.
The main task works from beginning to end
For inventory: product → receipt → issue → balance. For data: question → clean → analyse → explain.
You checked more than the happy path
Try empty inputs, duplicates, invalid values, unauthorised access, lost connectivity and other realistic failure cases.
Another person can inspect the work
Use the right evidence for the field: repository, demo, dashboard, notebook, lab report, screenshots or test results.
You can defend your decisions
Be ready to explain what failed, why you chose an approach, what you personally built and what you would improve.
Software development
Projects 1–3: full-stack development
These projects create evidence around application logic, databases, validation, APIs, authentication and testing. For the broader developer career route, see NIT’s How to Become a Software Developer in Kenya guide.
Small-business inventory and sales system
Build: products, stock receipts, stock issues/sales, calculated balance and low-stock view.
Test: if 40 units + 30 received − 22 sold = 48, then attempt to sell 60. The system should enforce the stock rule.
Publish: database structure, one complete transaction, validation tests, screenshots/demo and repository.
Proves: relational data, calculations, business rules, validation, testing and reporting.
Appointment and booking system
Build: services, available slots, booking form, status and a daily appointment view.
Test: create Tuesday 10:00 a.m., then try to book the same resource and slot again. Also test past dates and duplicate submission.
Publish: booking workflow, data model, conflict rule and tests.
Proves: dates/time, database relationships, validation and workflow logic.
Order and delivery-status tracker
Build: New → Confirmed → Preparing → Dispatched → Delivered, with status history and user permissions.
Test: try viewing another customer’s order, skipping required stages and changing a cancelled order.
Publish: status model, API/database records, access tests and a complete order journey.
Proves: authentication, authorisation, APIs, relational data and audit history.
Why these features matter
Data science & analytics
Projects 4–5: turn raw data into a decision
A data project should show the work between the raw file and the final chart. The Kenya National Bureau of Statistics provides official data portals that can support appropriately scoped public-data analysis. [4]
Kenya public-data analysis
Build: question → source data → cleaning → analysis → visualisation → interpretation.
Test: verify an important calculation independently and document missing values, duplicates or inconsistent formats.
Publish: question, data source, cleaning record, SQL/Python/Excel work, chart/dashboard, findings and limitations.
Proves: data quality, reproducibility, analysis and communication.
Small-business sales analysis
Build: analyse a fictional, open or authorised transaction dataset to answer which products drive sales and how performance changes over time.
Test: confirm totals, duplicates, date ranges, missing quantities and category reconciliation.
Publish: cleaning log, queries/code, dashboard, findings, recommendations and limitations.
Proves: SQL/Python/Excel, business reasoning, validation and visual communication.
Begin with a question. The visualisation is useful only when it helps answer that question or reveals the next useful question. For a Kenya-focused data project, official KNBS portals can provide credible datasets that you can clean, validate, analyse and explain. [4]
Cybersecurity
Projects 6–7: show evidence, investigation and verification
Login-monitoring laboratory
Build: in systems you own or an authorised lab, generate normal logins and deliberate failed-login events, then investigate account, time, source and frequency.
Test: recreate the event and confirm the detection still works when the account or number of failures changes.
Publish: lab diagram, sanitised logs, query/rule, investigation notes and final incident summary.
Proves: logging, detection, investigation and technical reporting. OWASP lists Security Logging & Alerting Failures in its 2025 Top 10 and specifically discusses login and failed-login events. [5]
Secure-application assessment and remediation
Build: use your own or an authorised training application. Complete scope → finding → evidence → correction → retest.
Test: for example, verify that an ordinary user can no longer reach an administrator-only function after the correction.
Publish: sanitised finding, remediation and retest evidence—never secrets or private system details.
Proves: security testing, access control, remediation and verification. Broken Access Control remains A01 in OWASP Top 10:2025. [6]
Artificial intelligence
Projects 8–9: evaluate the system, not just the demo
Document question-and-answer assistant
Build: public/non-confidential documents → retrieval → relevant source → answer → source display.
Test: easy factual question, multi-section question, ambiguous question, unanswered question and misleading question.
Publish: system diagram, source documents, retrieval process, test set, actual results and failure cases.
Proves: Python, text processing, retrieval, AI integration, evaluation and error analysis.
Prediction or classification project
Build: define target → inspect data → prepare data → create baseline → train → evaluate → inspect errors → explain limitations.
Test: look beyond one accuracy number. Check where the model fails and whether the chosen metric matches the problem.
Publish: dataset source, preparation, baseline, model, evaluation, failure analysis and next improvement.
Proves: modelling, evaluation, statistical reasoning and responsible interpretation.
Mobile application development
Projects 10–11: prove storage, validation and real device behaviour
Expense-tracking application
Build: amount, category, date and note; then monthly/category summaries, search, editing and local storage.
Test: KSh 200 + KSh 350 + KSh 100 should total KSh 650. Also try zero/negative values, empty categories, bad dates and app restart.
Publish: flow, storage method, screenshots/demo and test cases.
Proves: mobile UI, state, local persistence, validation and testing.
Offline event check-in application
Build: attendee list, unique code, check-in status/time and local storage that continues working when connectivity disappears.
Test: online check-ins → disconnect → continue → restart app → reconnect → synchronise → check for duplicates.
Publish: online/offline flow, persistence approach, conflict handling and synchronisation results.
Proves: offline-first design, local persistence, synchronisation and error handling.
Project 12 · Flagship
Build one flagship project for the exact job you want
Once you can finish smaller projects, collect several genuine vacancies for the role you want and look for requirements that repeat. Build one manageable system that makes several of those capabilities visible.
| Target role | Repeated requirements might include | Flagship direction |
|---|---|---|
| Backend / Full-stack | REST APIs, PostgreSQL, auth, Git, tests, deployment | Customer order system + API + role permissions + tests + deployment |
| Data analyst | Python, SQL, data validation, pipelines, dashboards | Ingestion → cleaning → database → quality checks → analysis → dashboard |
| AI developer | Python, model/LLM APIs, retrieval, evaluation, deployment | Evaluated AI application with source data, tests, failure analysis and deployment |
| Mobile developer | Flutter/Kotlin, APIs, storage, testing, release workflow | Mobile application + API + local persistence + error handling + device tests |
On a phone, swipe horizontally to compare the complete role, requirement and flagship-project columns.
Add a technology because it solves a problem or proves a relevant capability—not because its name makes the project sound more advanced.
Two practical examples
What a complete project workflow looks like
Example 1: from a shop notebook to an inventory system
Understand the process
Products exist, stock arrives, stock leaves and the shop needs a reliable balance. Use fictional records rather than copying private business data.
Build one transaction at a time
Finish Add Product. Test it. Then add Receipt. Test it. Then add Issue/Sale. Keep the scope controlled.
Prove the rule
If opening = 20, received = 40 and issued = 15, expected balance = 45. Then attempt to issue 60 and verify the system blocks the invalid transaction.
“I built an inventory application around stock transactions. The balance is calculated from recorded receipts and issues. I added validation to prevent issuing more stock than is available and tested the workflow after page reloads.”
Example 2: from a messy spreadsheet to a data project
Inspect and clean
Find missing names, inconsistent dates, duplicate transactions and inconsistent product labels. Record the corrections you make.
Analyse a question
Calculate monthly revenue, units sold, revenue by product and category. Build a visualisation that answers the original question.
Interpret, do not just chart
If a top product falls in September, ask whether stock, price, seasonality or substitution could explain it. State what the data can and cannot prove.
Repeatable method
Turn a job advert into a portfolio project
Do not design a project from one vacancy. Collect several realistic openings, then look for repeated requirements.
| Step | What to do | Output |
|---|---|---|
| 1 · Collect | Gather about five vacancies for the role you could realistically target. | A small sample of real requirements. |
| 2 · Count | Highlight skills or responsibilities that recur: SQL, APIs, Git, testing, authentication, dashboards, mobile storage. | A ranked requirement list. |
| 3 · Translate | Separate a skill name from evidence. “SQL” becomes schema, joins, constraints, queries and validation. | Inspectable evidence for each skill. |
| 4 · Scope | Choose one manageable problem. Avoid building a giant ERP just to include every technology. | A project you can finish. |
| 5 · Prove | Build the evidence deliberately and test important failure cases. | A job-targeted portfolio project. |
On a phone, swipe horizontally to view the full five-step job-advert workflow table.
Present the work clearly
Turn every major project into an eight-part evidence pack
Problem + intended user
Explain what the project solves and who needs it. Keep both statements short.
Your contribution + working evidence
State what you personally built, then provide the best demo form for the field: live app, dashboard, recording, report or lab output.
Technical evidence + tests
Where appropriate show source, schema, SQL, architecture, test files and expected-versus-actual results.
Challenge + next improvement
Describe one real failure, correction and retest. Then explain what you would improve next.
For code projects, GitHub recommends useful READMEs with key features, setup instructions, a demo/example and tests. [1] Its README guidance also describes the README as a place to explain what the project does, why it is useful and how to get started. [2]
Fast review
Use the two-minute reviewer test
GitHub notes that hiring managers may spend only a couple of minutes looking at projects. [1] Open your project as though you have never seen it before. Can you quickly find the problem, intended user, working result, technologies, your contribution, tests and demonstration?
A technically strong project can still make weak portfolio evidence when the reviewer has to hunt through folders or screenshots to understand what happened.
Quality check
Use this 100-point project scorecard before publishing
| Area | Question | Score |
|---|---|---|
| Job relevance | Does the project demonstrate something the target role actually requests? | 0–20 |
| Completeness | Does the main workflow work from beginning to end? | 0–20 |
| Testing | Did you test important valid, invalid and failure cases? | 0–20 |
| Evidence | Can another person inspect what you built and how it works? | 0–20 |
| Ownership | Can you explain decisions, errors, fixes and your own contribution? | 0–20 |
On a phone, swipe horizontally to view the complete scorecard. Maximum: 100. The number is not a qualification. Use the weakest category to decide what to improve next.
Develop the work you already have
Upgrade a beginner project instead of constantly starting again
Start simple
Exercise → working feature
An expense tracker may begin with one form that stores a transaction. A to-do list may begin with one user and one status.
Add depth
Feature → complete workflow
Add persistence, validation, users, reports, tests, documentation and deployment only as your understanding grows.
Your portfolio becomes stronger when you stop following the tutorial line by line and begin making your own requirements, data, rules, tests and design decisions.
Professional judgement
Use school, attachment and AI-assisted work safely
School and group projects
A school project can belong in your portfolio. Explain the assignment, your personal contribution, what teammates handled, how the work was tested and what you learnt. Do not claim another person’s contribution as your own.
Industrial attachment work
Working on an organisation’s system does not automatically give you permission to publish its source code, customer records, database exports or internal diagrams. Recreate the technical idea using fictional information where necessary. NIT’s ICT Industrial Attachment in Kenya guide covers workplace evidence and confidentiality in more detail.
AI-assisted projects
AI can assist with code or analysis, but you should still read, run, test, change and explain the result. Be ready to state what AI helped generate, what you changed and how you verified the output.
Passwords, API keys and personal data
Do not publish passwords, tokens, database credentials or private records. GitHub warns that sensitive data committed to a repository can remain in Git history even after it is removed from the latest version. [3]
For student projects, prefer fictional, anonymised, authorised or properly public data. Kenya’s Office of the Data Protection Commissioner sets out principles including lawful, fair and transparent processing, purpose limitation and data minimisation. [7]
Before you apply
Run one final inspection of every project
Open every link. Run the main workflow. Check the README. Remove secrets. Confirm the demo still works. Then practise explaining the problem, user, your contribution, one difficult failure, how you tested it and the next improvement.
The final question: can you point to at least one important requirement in the target role and show exactly where your project demonstrates it?
NIT’s current ICT programme guide lists Full-Stack Software Development, Cybersecurity, Data Science & Analytics, Artificial Intelligence and Mobile App Development with practical project work and compulsory industrial attachment. [8]
Quick reference
ICT portfolio projects in Kenya: frequently asked questions
Which ICT projects are best for getting a job in Kenya?
The best project depends on the role. Full-stack candidates can build complete database applications; data candidates can show end-to-end analyses; cybersecurity candidates can show authorised labs and technical reports; AI candidates can show evaluated AI systems; and mobile developers can demonstrate working apps with storage, validation and testing.
How do I choose a portfolio project?
Read several realistic vacancies for the role you want. Identify repeated requirements, then choose a manageable project that creates visible evidence of those skills.
Do I need twelve projects?
No. The twelve projects here are choices. Three strong projects are enough to begin: one foundation project, one role-specific project and one flagship project. GitHub recommends highlighting roughly three to five relevant projects. [1]
Do I need GitHub for an ICT portfolio?
GitHub is particularly useful for code-based software, data and AI work. Other evidence may be more suitable for some tasks, such as dashboards, technical reports, lab outputs, recorded demonstrations or released mobile applications.
Can personal projects help if I have no job experience?
Should every software project be deployed?
A working deployment is useful because a reviewer can try the result without configuring it locally. Where public deployment is impractical, provide a clear demonstration and reliable setup instructions.
Can a copied tutorial project go in my portfolio?
Use tutorials to learn, then change the requirements, data, workflow, rules or tests until you have made meaningful technical decisions yourself. You should be able to explain the final version without following the tutorial.
Can I use AI to help build a project?
Yes, but you remain responsible for understanding and testing the work you present. Read generated code, run it, change it, test failure cases and be able to explain the result.
Build skills you can demonstrate
Register for an ICT Course at Newton Institute of Technology
Train in Full-Stack Software Development, Cybersecurity, Data Science & Analytics, Artificial Intelligence or Mobile App Development through practical learning, project work and compulsory industrial attachment.
Sources and references
- GitHub Docs. Using your GitHub profile to enhance your resume. Guidance on highlighting 3–5 relevant projects, READMEs, demos and tests for hiring review. Accessed 27 October 2026. docs.github.com.
- GitHub Docs. About READMEs. Repository documentation guidance. Accessed 27 October 2026. docs.github.com.
- GitHub Docs. Removing sensitive data from a repository. Guidance on secrets and Git history. Accessed 27 October 2026. docs.github.com.
- Kenya National Bureau of Statistics. KNBS Data Portals. Official Kenyan statistical data resources. Accessed 27 October 2026. knbs.or.ke.
- OWASP Foundation. A09:2025 Security Logging & Alerting Failures. Accessed 27 October 2026. top10.owasp.org.
- OWASP Foundation. A01:2025 Broken Access Control. Accessed 27 October 2026. top10.owasp.org.
- Office of the Data Protection Commissioner, Kenya. Rights of a Data Subject / Principles of Data Protection. Accessed 27 October 2026. odpc.go.ke.
- Newton Institute of Technology. ICT Courses in Kenya: Fees, Requirements & Duration. Current NIT programme structure for Full-Stack Software Development, Cybersecurity, Data Science & Analytics, Artificial Intelligence and Mobile App Development, including industrial attachment. Accessed 27 October 2026. niteducation.com.
Prepared by Newton Institute of Technology Editorial Team · Technical review: Engineer Simon Barongo · Last reviewed: 27 October 2026.