Data is produced at scale across every African institution — mobile-money rails, microcredit ledgers, agricultural cooperatives, health systems. Most of it is never analyzed. When it is, the tools and benchmarks come from elsewhere, calibrated on data that looks nothing like the population the institution actually serves.
The Applied AI & Data Science for African Systems program addresses both gaps: it builds a rigorous, end-to-end working capability in modern applied AI and data science — and it does so on case studies, datasets, and regulatory contexts drawn from West Africa. Curriculum delivered by AIRINA Labs research faculty, with industry mentors from banks, MFIs, telecoms, and the African research diaspora.
Program Overview
Over 14 weeks, participants move from foundations to a defended capstone project. The program is taught live online — bilingual, cohort-paced, mentored — with weekly research-led lectures by AIRINA faculty and weekend working sessions with industry mentors. Course materials, code, and reading lists are open.
Unlike a generic data-science MOOC, every module is anchored in West African institutional data realities: credit scoring on BCEAO-zone microcredit, anomaly detection on mobile-money transaction streams, regulatory reporting under BCEAO/BCBS interpretability constraints, and applied generative AI for institutional knowledge work in French and English.
Program structure
- Pre-work (2 weeks). Python for data science, statistics refresher, the data science lifecycle, the evolution of AI through Generative AI — with practice quizzes. So the cohort arrives at the live sessions on the same footing.
- Core curriculum (8 weeks). Live, research-led, working from real institutional case studies. Detail in the Learning Outcomes tab.
- Project window (1 week). Time to finalize and submit individual mid-program projects.
- Capstone (3 weeks). Integrative team capstone on a real institutional dataset (under NDA where required). Final defense before AIRINA faculty and invited industry reviewers.
Mentorship model
Beyond the lecture stream, every participant is assigned to a small mentored micro-group (5–7 people grouped by background) led by an industry mentor — senior data scientists, ML engineers, and applied-AI practitioners drawn from AIRINA's partner network. Mentors meet the group weekly to translate concepts into hands-on practice, run coding walkthroughs, and prepare participants for the capstone defense.
Certificate
Successful completion — pre-work + core modules + capstone defense — earns an AIRINA Labs certificate in Applied AI & Data Science, graded, with the capstone reviewer's comments included.
Learning Outcomes
By the end of the program, participants will be able to:
- Frame an applied AI & data science problem from a real West African institutional context — translating an operational question into a data-science scoping document defensible to both engineering and regulatory reviewers.
- Build the standard supervised, unsupervised, and time-series ML pipeline end-to-end in Python (NumPy / pandas / scikit-learn / PyTorch) on real institutional data.
- Design and audit interpretable models for high-stakes financial-inclusion decisions — GAMs, monotonic gradient boosting, SHAP/LIME, counterfactuals, and fairness audits under BCEAO regulatory constraints.
- Apply modern deep learning (CNNs, encoder-decoder architectures, attention, transformer-based methods) to text, image, and time-series problems originating from African systems.
- Use modern generative AI and agentic AI as a working tool — prompt engineering, retrieval-augmented generation, light agent construction — with explicit awareness of failure modes and ethical risks.
- Deliver a defended capstone: scoped problem, dataset, model, evaluation, fairness audit, written report, oral defense.
Program curriculum (weeks 1–14)
- Python for data science (NumPy, pandas)
- Python for visualization (Matplotlib, Plotly)
- Inferential statistics
- Hypothesis testing
- Hypothesis testing in applied contexts
- Dimensionality reduction: PCA, t-SNE
- Network analysis on transaction data
- Clustering algorithms and their failure modes
- Maximum likelihood, Bayesian estimators
- Linear regression & its assumptions
- Cross-validation, bootstrapping
- Logistic regression, KNN, Gaussian models
- Decision trees, entropy & information gain
- Ensemble learning: bagging, random forests, gradient boosting
- Time-series forecasting on financial-inclusion data
- From perceptron to multi-layer architectures
- Convolutions, pooling, CNN architectures
- Transfer learning & augmentation
- Encoder-decoder, attention mechanism, positional encoding
- Content-based recommendation
- Collaborative filtering & SVT
- Matrix estimation methods
Time to finalize and submit mid-program projects. Mentor review and feedback.
- Origins of generative modeling
- Generative AI as a matrix-estimation problem
- LLMs as probabilistic models for sequence completion
- Prompt engineering — practice and pitfalls
- Summarization, classification, generation
- Retrieval-augmented generation (RAG)
- Agentic AI — light agents and their failure modes
- Bilingual EN/FR application contexts
Integrative team capstone on a real institutional dataset (under NDA where required). Final defense before AIRINA faculty and invited industry reviewers. Written report graded; cohort presentation.
Who Should Attend
This program is for working and aspiring data & AI professionals who want a rigorous end-to-end working capability, calibrated to West African institutional realities.
- Data scientists, ML engineers, and analysts in West African banks, microfinance institutions, mobile-money operators, fintechs, and insurers.
- Engineers and software practitioners transitioning into applied AI & data science roles.
- Graduate students in mathematics, statistics, computer science, or quantitative economics — including AIMS alumni and partner-university students.
- Public-sector analysts and regulatory-affairs professionals working with quantitative methods in West Africa.
- Entrepreneurs building data- or AI-driven products for African markets.
Prerequisites
- Programming. Working Python (functions, classes, NumPy/pandas at the level of small data manipulations).
- Mathematics. Calculus and linear algebra at undergraduate level; comfort with basic probability and statistics.
- Time commitment. ~8–10 hours per week across 14 weeks — live lectures (2–3 hrs), mentored sessions (1–2 hrs), and individual project work (4–6 hrs).
Selection
For oversubscribed cohorts, applicants complete a one-question selection prompt drawn from their stated background. Selection priority is given to applicants from the BCEAO zone and to underrepresented groups in technical fields.
Brochure
The detailed program brochure (PDF, EN/FR) is sent on request — including week-by-week curriculum, mentor profiles, capstone examples, and the cohort calendar.
To receive the current brochure, write to contact@airina.africa with "Applied AI & DS — brochure request" in the subject. The brochure is updated each cohort; we send the version current at the time of your request.