Research & Development

What we're building:Authorship analysis for communications written in Swahili, Sheng, and English code-switching, developed as decision-support for forensic examiners.

Why it matters:Many authorship attribution tools are built for single-language English text. Disputed chats, extortion messages, and threat communications in East Africa often mix languages.

What we're building:Research into detecting artificial intelligence generated audio, video, and image artefacts featuring African speakers, accents, and diverse skin tones.

Why it matters:Public detection models are often trained on little African data and can perform worse on African voices and faces.

What we're building:Graph analysis of mobile-money and bank statements supplied directly by clients, to trace how disputed funds move through account networks.

Why it matters:Fraudulent transaction chains can spread value across many mobile wallets quickly. Mapping these trails supports recovery and dispute reporting.

What we're building:Hash-chained, time-stamped chain-of-custody logging records, ensuring any subsequent modification to an evidence custody record is mathematically detectable.

Why it matters:Designed to provide integrity documentation that counsel, courts, and opposing examiners could inspect independently to confirm evidence history.

What we're building:Protecting long-term evidence archives by applying NIST post-quantum digital signatures alongside contemporary cryptographic hashes.

Why it matters:Forensic case records should stay verifiable for decades. Quantum-resilient integrity signatures are designed to keep archived records verifiable as computing power grows.

What we're building:Developing better methods for visualising latent fingerprints on thermal receipts without blackening or damaging the paper.

Why it matters:Mobile-money agent slips, ATM receipts, and point-of-sale printouts can be important physical exhibits in commercial disputes, but conventional development techniques can darken or destroy them.