A short account of who we are, the markets we serve, and the standards to which we hold our own work.
Geoinfortrak International Limited is a Kenyan technology company registered under Certificate Number PVT-ZQU5K56, headquartered in Nairobi, and engaging clients across three complementary markets: the African continent, the Gulf Co-operation Council, and the global African and Arab diaspora.
We were founded on a single observation: the technology industry that serves these markets has too often been imported wholesale from outside them, then fitted imperfectly to local realities. The remedy, we believe, is not to complain of this — but to build the alternative, discipline by discipline, product by product, engagement by engagement. Where an engagement calls for domicile, accreditation, or sector credentials we do not hold ourselves, we co-deliver with trusted regional partners rather than overreach.
Our practice spans six interdependent disciplines. An artificial-intelligence system built without geospatial awareness misses half the territory; an electronic commerce platform built without rigorous software engineering becomes shelfware within two fiscal years. We treat these disciplines as a single practice because they are, in fact, a single practice.
We would rather deliver one product that functions in production than pitch ten that do not.
Beautiful architecture that never ships is of less value than imperfect code running in production. We build iteratively, measure against reality, and refactor on evidence rather than on taste.
M-Pesa and mada before Stripe. Swahili and Arabic before English-only. Counties, ministries, and emirates before capitals elsewhere. We default to the payment rails, languages, and institutional forms that actually exist in the markets we serve — not the ones that exist in the decks that reach us from outside them.
We document what is fragile. We declare what is covering for what. We prefer to deliver an unvarnished assessment of a system's condition than an impressive-sounding account that reality will later contradict.
This principle governs our geospatial, AI, and software work alike. Models are instruments, not verdicts. We maintain scepticism of our own outputs so that our clients are not required to.
A geospatial and artificial-intelligence engineer with more than fifteen years of applied experience across infrastructure, utilities, and the built environment. Holds a Master of Science in Artificial Intelligence from the University of Leeds and a Bachelor of Technology in Geoinformation Technology from the Technical University of Kenya.
Core capabilities span the integration of Building Information Modelling with enterprise geographic information systems; LiDAR-derived feature extraction and automated asset classification using deep-learning methods including PointCNN, U-Net, and Mask R-CNN; enterprise asset management for roads, drainage, and utility networks; geospatial data re-engineering and schema migration at national scale; and the design, evaluation, and deployment of machine-learning systems, including large-language-model integration and agentic workflows. Fluent across the ESRI and ArcGIS platform, open-source geospatial tooling, and the modern Python data-science and deep-learning ecosystem.
Works at the intersection of engineering rigour and institutional pragmatism — directing Geoinfortrak's consultancy engagements, the AI Champs product line, and selective advisory mandates across Africa, the GCC, and the diaspora.