From computer vision and personal finance to institutional sports and streaming ML — real systems, shipped and running, built for Tanzanian realities.
An academic road-damage scanner built by Afritis. It detects longitudinal, transverse, and alligator cracks, plus potholes — then shows confidence for each finding. Compare training checkpoints on the same photo. Available live for demos and research exploration.
A personal-finance operating system designed around a simple reality: many people use one bank account but need several financial purposes. WealthFlow turns one real account into clear virtual wallets — tracking income, expenses, debts, and goals in a TZS-native, mobile-friendly experience.
A mobile and web platform for Shirikisho la Michezo Taasisi za Umma na Binafsi Tanzania — digitising how institutions manage athletes, teams, fixtures, results, standings, and daily sports administration.
A streaming machine-learning pipeline for market sentiment — Kafka producers, Spark Structured Streaming, model inference via FastAPI, persistent storage, live dashboards, and containerised deployment.
A platform concept connecting customers, businesses, and professionals through location-aware discovery, provider profiles, service pricing, emergency availability, and AI-supported recommendations.
More engineering work and experiments live on GitHub.
Browse GitHub ↗Early research directions we are exploring — problem spaces we want to grow into, described honestly as interests rather than finished work.
How sensing, vision, and decision-support could help smallholder farming — from crop monitoring to practical field insight.
Digital-health and AI-assisted tools that make information, screening, and care workflows clearer in low-resource settings.
Practical NLP, retrieval-augmented assistants, and interfaces that make complex information easier to use in local languages.
We take projects from first conversation to working product. Tell us what you are trying to solve.
Start a project →