Systems in the field, described plainly.

Each entry states the problem we were handed, the approach we took, and what the system does now. Where a result is still being measured, we say so rather than publish a number we cannot stand behind.

Recognise your problem in one of these?

Most of what we build starts as someone describing a spreadsheet that stopped scaling.

Related work

Umuseke voice advisory

A Kinyarwanda speech model that turns a field voice note into a structured agronomic query and answer.