AIRINA is a research institute in Cotonou, founded in 2025. We work on topology and applied machine learning, applied to the questions West African banks and microfinance institutions actually ask. We publish before we claim. The institutions that pay our bills are the ones that send us their problems; the trainees do not pay.
Cotonou, Bénin · Phase 1 of 3 — finance, agriculture, health.
Looking for credit scoring, anomaly detection, or risk tools deployable in the BCEAO zone.
See the products →Interested in topology, persistent homology, or interpretable ML applied to African data.
See the research →Mathematician, statistician, or engineer ready for short-form, graded, research-led programs.
See the catalog →An institution that wants to fund a cohort, sponsor a research theme, or build with us.
Get in touch →Mostly at the intersection of topology and applied machine learning. Persistent homology, interpretable models, asymmetric topology, a small bet on quantum-topological methods for finance. The papers are on arXiv; the code is on GitHub. Six active themes →
Research turned into things banks and microfinance institutions can deploy. Right now: credit scoring and anomaly detection for institutions in the BCEAO zone. Later: agricultural risk. Later still: health. How we build →
Short cohorts, taught by the people doing the research. Two days to two weeks. Every program ends with a graded defense. Bilingual delivery. Philosophy → · Catalog →
AIRINA Ventures opens in year three. Not active yet. Written down on this page so we hold ourselves to it. What it will be →
Methods paper targeting Q3 2026 submission. Companion code on GitHub once accepted.
Three days, in-person Cotonou, ~12 participants. Hands-on with GUDHI, Ripser, giotto-tda.
Partner under NDA. Methods will be published on GitHub at the end of the engagement.
A reading group on topological data analysis, interpretable ML, and applied mathematics for African systems. One paper, two discussants, an hour of careful argument. Read it if you can; come anyway if you can't — everyone is welcome, including first-timers. Hybrid: in-person at the Cotonou office or by Zoom.
The canonical introduction to TDA. Every subsequent paper assumes you have read this one.
The founding algorithm of persistent homology. Where the field was named and proven.
The most beautiful applied-TDA paper: a Klein bottle hiding in the patches of every photograph.
Mapper, the second pillar of TDA. Required before anything Mapper-shaped shows up in our own work.
The bridge from persistence diagrams to standard ML pipelines. Required before any downstream classifier reads a TDA feature.
The literature directly under our Phase 1 products. We agree, we disagree, we read it carefully.
The stability theorem. Why persistent homology is a real signal and not noise.
The other major vectorization of persistence diagrams. Bubenik vs. Adams, head to head.
The case for interpretable models, not post-hoc interpretations of black boxes. Why AIRINA products must be the former.
The high-stakes-medicine TDA paper. Precedent for our Phase 3 health work.
Six programs across three tiers, with two anchor cohorts opening Q3 and Q4 2026. See the catalog · version française.
Three days, in-person Cotonou, ~12 participants. Hands-on with GUDHI, Ripser, giotto-tda. Full curriculum.
Phase 1 of three: applied AI for financial inclusion in West and Central Africa. Our research themes.
We publish before we claim. We say so when we don't know.
The questions we work on come from African institutions. Benchmarks from elsewhere do not earn a pass.
We exist for the people the formal economy has excluded. If our work cannot reach them, it is not the work we want to do.
We are building an institution, not a project. The decade matters more than the quarter.
Mathematician working at the intersection of topology, geometry, and applied AI. Research interests include topological data analysis, persistent homology, fixed-point theory, and machine learning for African systems.
Mathematician and frequent research collaborator with Y. U. Gaba on topology and applied analysis.
Cotonou
Benin Republic
Tuesdays 14:00–16:00 WAT
Walk-in or by Zoom