PhD Engineering researcher, Imperial College London (Shawki Saad Scholar), on the VIEW project, funded by Schmidt Sciences.
I work on AI for environmental science: how far a measurement can be trusted, and what it is worth to a decision. Fault detection on low-cost water sensors, flood warning where ground truth is scarce, and AI systems that respect the physics and provenance of their data.
Science · Build · Deploy in Africa
- Reliability-Caps: the value of a single resident's flood report to a spatial decision.
- leakage-benchmarks: code for Removing Data Leakage Does Not Fix Benchmarks.
- GageLink: hydrology data for AI agents, with units, datums and record quality attached.
- Talk: Mapping Floods Where There Is No Ground Truth, London Geo Meetup, Birkbeck, 2026.
- ng-flood: Google flood forecasts for any point in Nigeria, by postcode area.
- ng-postcode: Nigeria's NIPOST digital postcode in Rust, Python, JavaScript, Java and PostgreSQL.
- WazoCore: Yoruba, Hausa and Igbo in a third fewer tokens (model on Hugging Face).
Head of Research & Innovation, Geohazards Risk Mapping Initiative (GRMI), Nigeria. GeoAI flood mapping for 25 communities; citizen flood reports to NEMA; flood-risk training and flood-tolerant seed for rice farmers in Katcha Ward, 2026 (FairPlanet, MSME Africa).
- PhD Engineering, Imperial College London (Shawki Saad Scholarship)
- MSc, London School of Economics, Distinction (full merit scholarship)
- BSc Geography, University of Ilorin, best graduating student (full merit scholarship)
Flutterwave, software engineer · UNDP, AI policy
freeCodeCamp (9 tutorials) · LogRocket (60+ articles on building AI products)