Last checked
Most data careers in Kenya start at data analyst: Excel, SQL and Power BI on a bank's, telco's or NGO's records. Python, statistics and machine learning then open data scientist, ML engineer and AI engineer roles. Self-taught, university and bootcamp routes all work. Interviews turn on a portfolio built on African data that you can explain.
McTaba runs coding and data bootcamps in Kenya. This guide is written to be useful whether or not you ever join one.
Our founder led software engineering at Zindua School from January 2023 to February 2024.
Open the M-Pesa app and request a statement. The PDF that lands in your email lists every shilling you moved, with the date, the other party and the balance after. A data scientist's work starts with files like that one.
Kenyan organisations hold piles of them. A SACCO keeps a loan ledger and a list of guarantors. A hospital keeps a patient register. A duka keeps a till roll, and a chama keeps contributions in an exercise book or a WhatsApp group. Someone has to turn those records into a decision.
This map follows one person from a first line of Python to a first job. It names five roles and the skills in the order you meet them. Then come the sectors that hire, three routes in, and the portfolio that earns an interview. Every guide in this hub is listed at the end, by stage.
Who wrote it: Mctaba Labs, a Nairobi software and AI studio that also trains. We build on M-Pesa Daraja, USSD and the WhatsApp Business API, so our examples use that kind of data.
Job titles blur in Kenya. One bank's "data analyst" builds credit models while another's refreshes an Excel sheet every Monday morning. Read the duties in the advert, not the title.
Here is what each role hands over at the end of a working week.
| Role | The question on your desk | Tools you open every day | What you hand over |
|---|---|---|---|
| Data analyst | "Why did deposits at the Kisumu branch drop in March?" | Excel, SQL, Power BI | A chart and three sentences a manager can act on |
| BI analyst | "Build the weekly sales pack for every branch, and make it refresh itself." | SQL, Power BI or Tableau, a data warehouse | A dashboard that is ready before the 8 AM meeting |
| Data scientist | "Which chama members will miss next month's contribution?" | Python, pandas, scikit-learn, statistics | A model, its error rate on data it has never seen, and the cost of each mistake |
| ML engineer | "The credit model works in a notebook. Now it must answer every loan request in under a second." | Python, Docker, FastAPI, a cloud platform | A model behind an API, with alerts when its inputs drift |
| AI engineer | "Let customers ask for their loan balance on WhatsApp." | LLM APIs, retrieval, evaluation sets | An assistant that answers from the company's own records and refuses what it should |
The ladder runs in that order because each role uses the one below it. An AI engineer still writes SQL. Our Data Science to AI Engineering programme climbs the same ladder, from Python in week one to an AI engineering phase near the end. You can climb it on your own too, and the rest of this map shows how.
A sixth title appears in many adverts: data engineer, who builds the pipes that carry data to everyone else. Read how a data analyst, a data scientist and a data engineer differ in Kenya. For the bigger labels, read what separates AI, machine learning and data science.
Get weekly developer tips
Join 25,000+ developers. Practical guides, job tips, and new content — straight to your inbox.
No spam. Unsubscribe anytime.
Many beginners learn data skills in the wrong order. They start with neural networks on YouTube, then freeze when an interviewer asks for a SQL join. The order below follows how the work arrives at a Kenyan desk.
Maths sits under steps four and six, and you do not need it on day one. If school maths scares you, read whether you can learn data science without a maths background.
Two questions come up before step three. Should you learn Power BI or Python first? That depends on the adverts you are aiming at: whether Kenyan data jobs ask for Power BI or Python. And what should each month look like? The week-by-week plan lives in the data science roadmap for beginners.
On BrighterMonday, "data analyst" returned 41 adverts and "data scientist" returned 17 (BrighterMonday data analyst search, BrighterMonday data scientist search, checked on 3 November 2026). Analyst roles outnumber scientist roles, which is why most people enter as analysts.
The employers behind those adverts fall into a handful of sectors. Each holds its own kind of data and asks its own kind of question.
Loan books, card swipes, KYC scans and claims files. The questions: who will default, which claim is padded, which customer is about to move to another bank. Co-operative Bank advertised data scientist, data architect and AI engineer roles. Britam wanted a forensic data analyst (BrighterMonday, MyJobMag, checked on 3 November 2026).
Call records, airtime top-ups, bundle purchases and the float sitting with every M-Pesa agent. The questions: which customers will switch networks, which agent runs out of cash before the Friday rush, which bundle price sells. Telco data feeds other employers too. GiveDirectly advertised a senior data scientist in Kisumu for mobile network data (BrighterMonday, checked on 3 November 2026).
Mobile loan applications, repayment histories and phone metadata. The questions: how much to lend a first-time borrower and how to spot a fraud ring. Pezesha advertised a junior data scientist and Watu Credit a data engineer (BrighterMonday, checked on 3 November 2026). Fraud work has its own guide: how machine learning detects mobile money fraud.
Member contributions, guarantor lists and dividend books. Many SACCOs size a loan by the member's savings and back it with guarantors' signatures. A model that reads two years of contributions can add a second opinion: how to build a credit score for chama members.
Household surveys, cash transfer records and field trial results, often collected on phones in remote counties. GiveDirectly, the International Rescue Committee and ILRI all had data or analyst adverts open (BrighterMonday, checked on 3 November 2026). NGO work pays attention to method, so statistics matter more here than flashy models.
Patient registers, lab results, stock cards for medicines and records from community health workers. The questions: which clinic will run out of malaria tests, which mothers miss their second antenatal visit. Read how data science is used in Kenyan healthcare.
Satellite and drone images, weather records, farm input sales and milk collection logs. Victory Farms advertised a drone data analyst, Pula a remote sensing data analyst and SunCulture a data engineer (MyJobMag, BrighterMonday, checked on 3 November 2026). The full picture is in how data science is used in Kenyan agriculture.
The list is not closed. Kenya Airways, Weetabix and Mondelez also had data adverts listed that day (BrighterMonday, MyJobMag, checked on 3 November 2026). For where adverts appear and how to read them, see where data jobs in Kenya are advertised. When one turns into an interview, prepare with the questions Kenyan data interviews ask.
Most Kenyan data adverts hide the salary. Every data analyst and data scientist advert on the first page of BrighterMonday's results showed "Confidential" (BrighterMonday, checked on 3 November 2026).
Two sources give a rough frame. KNBS puts the 2025 average for private sector staff in information and communication at KSh 1,438,060, about KSh 119,838 a month (KNBS Economic Survey 2026, Table 3.7, checked on 3 November 2026). That average covers everyone in the industry, from junior technicians to chief technology officers. Treat it as context, not as an offer.
Paylab says 80% of data scientists in Kenya earn KES 51,664 to KES 195,546 a month before tax (Paylab Kenya, data scientist, checked on 3 November 2026). Its sample is fewer than 20 verified respondents, so one new salary can swing it.
For figures by level, sector and employer, read what a data analyst earns in Kenya and what a data scientist earns in Kenya. Whether a course pays back on those numbers is its own sum: is a data science course worth it in Kenya.
One rule holds across every route. No school, course or mentor can promise you a job or a salary. Treat any offer that does as a warning sign.
All three routes produce working data people in Kenya. They differ in cash, speed, and who checks your work before an employer does.
The cheapest route in cash. ALX's All Access plan costs $5 a month or $49 a year. It includes ALX's Data Science programme, listed at 22 weeks of 20 hours (ALX All Access, ALX Data Science, checked on 3 November 2026). DataCamp's free tier opens the first chapter of every course, and Premium costs $13 a month billed yearly (DataCamp pricing, checked on 3 November 2026).
The cost is hidden. Nobody marks your work, so a wrong SQL join can survive until an interviewer finds it. If you go this way, join a study group or find a working analyst who will read your code every week.
A degree is slow, but banks, government bodies and international NGOs often screen CVs by qualification. Strathmore's MSc in Data Science and Analytics runs two academic years of in-person classes, 5:30 to 8:30 PM on weekdays. Its programme page does not publish the fee (Strathmore University, checked on 3 November 2026). The Open University of Kenya lists its online Master of Data Science at KES 187,500 over two years (Open University of Kenya, checked on 3 November 2026). Both expect a first degree.
A bootcamp squeezes the skills into months, with deadlines, a cohort and someone who reviews your projects. Moringa School lists its data science bootcamp at KES 174,000 to KES 200,000. It runs 31 weeks full time or 50 weeks part time (Moringa School, checked on 3 November 2026). Zindua School lists its data science course at KES 130,000 paid at once, or KES 150,000 over five instalments. It runs 5 months full time or 8 months part time (Zindua School, checked on 3 November 2026).
Ours costs KES 140,000 and takes 5 months full time or 8 months part time. The next cohort starts on 11 January 2027. No coding, data or maths background needed.
The trade-off is cash and hours. A full-time track leaves little room for a job, and a part-time track stretches past half a year. Compare providers in which data science courses in Kenya suit which learner. Weighing Moringa? Read what the alternatives to Moringa's data science bootcamp are. If you have a job to keep, start with how to learn data science while working.
Still torn between data and code? Read whether to learn data science or software engineering before you pay anyone.
Beginner portfolios lean on the same tutorial datasets: the Titanic passenger list, Iris flowers, house prices from another continent. A reviewer who has seen them learns nothing new about you.
A portfolio built on African data shows you can handle the mess Kenyan employers own. Phone numbers typed three ways. Amounts with "Ksh" glued on. Two date formats in one column. Five projects that prove it:
Then ship each one. Put the model behind an API, a dashboard or a WhatsApp reply that a stranger can try. A notebook proves you can follow a tutorial. A link that answers proves you can finish.
Two cautions. Real customer records fall under the Data Protection Act, 2019. Strip names, phone numbers and ID numbers before anything leaves the office: how to anonymise personal data in Kenya. And public competitions sharpen your modelling against strangers: how a beginner should start on Kaggle.
For a longer list of ideas, and what each one proves to an employer, read which portfolio projects get a data scientist hired.
Time depends on hours per week far more than on talent. Our full-time plan makes a fair yardstick because it budgets every phase in weeks. Here it is, phase by phase:
| Weeks (full time) | Phase | What it covers |
|---|---|---|
| Weeks 1 to 2 | Intro to Data Science and Python | Python from zero, Git and GitHub, APIs, scraping, regex and tests |
| Weeks 3 to 6 | Data Handling, EDA and Business Intelligence | OOP, SQL, Pandas, data cleaning, EDA, Power BI and Streamlit |
| Weeks 7 to 8 | Math, Statistics and Time Series | Linear algebra and calculus through code, gradient descent from scratch, Bayes for fraud alerts, A/B testing and time features |
| Weeks 9 to 13 | Machine Learning | Linear and logistic models, metrics for imbalanced fraud data, XGBoost and LightGBM, clustering, SHAP and fairness |
| Weeks 14 to 16 | Deep Learning | Neural networks in PyTorch, CNNs and OCR for KYC documents, forecasting, NLP and fine tuning a transformer on Hugging Face |
| Weeks 17 to 18 | MLOps and Production AI | MLflow, Docker, FastAPI, CI/CD, Cloud Run, monitoring for drift, cloud pricing and fine tuning with LoRA |
| Weeks 19 to 21 | Real Time and Online Machine Learning | Kafka streams, concept drift, online learning with River, Vowpal Wabbit, bandits and drift detectors |
| Week 22 | Intro to AI Engineering | LLM APIs, prompting, retrieval augmented generation, agents, evaluation sets and guardrails |
| Weeks 23 to 24 | Capstone and Career | An end to end financial ML product built with a team, deployed and shown at a public demo day |
Part time, the same plan takes 8 months, with four sessions a week, 7 PM to 9 PM. On your own, the arithmetic is plain: the plan holds 288 contact hours, so at 10 hours a week the teaching alone fills 29 weeks. Projects, revision and the weeks when life gets in the way come on top.
Read the table as milestones. By the end of the second phase you can do a junior analyst's work: SQL, pandas and a Power BI dashboard. That is the point at which many people start applying, while they keep learning the later phases at night.
The job search is a phase of its own. Plan for applications, take-home tests and interviews, and keep shipping small projects while you wait. A fresh project on your GitHub gives you something new to talk about in every interview.
We run a data science programme, so read this section with that in mind. Several people are better served elsewhere.
Choose us if you want every model shipped behind an API, a dashboard or a WhatsApp reply, on African data. Capstone options include M-Pesa fraud detection, Credit scoring, Customer segmentation, RAG support assistant. Delivery is hybrid: live online sessions on weekdays and in person sessions in Nairobi on Saturdays, all recorded.
Each guide answers one question. Pick the stage you are at and start there.
Pick one dataset you can touch today. Your own M-Pesa statement works. So does a month of your chama's contributions, or the sales book from a relative's shop.
Load it into Google Sheets. Write down one question it can answer, such as "which week of the month do I spend most?" Answer it with a pivot table and one chart. Then do the same thing again next week in Python.
That small loop of question, data, chart and sentence is the whole job at a smaller size. If it holds your attention for a month, the rest of this map is worth your time.
Want a cohort, deadlines and facilitators who are Google Cloud Certified Machine Learning Engineers? Look at our Data Science to AI Engineering programme. It takes 5 months full time or 8 months part time, and the next cohort starts on 11 January 2027.
KES 140,000
Data Science to AI Engineering: 5 months full time or 8 months part time, hybrid in Nairobi. Fraud, credit and M-Pesa data from Week 1, and every model deployed. Starts 11 January 2027.
See the Data Science programme