Mosop
the highlands that calibrate the model.
Mosop is the Nandi word for the highlands — the elevated plateau where Uasin Gishu County rises into the Rift Valley escarpment. It is also where RaDa Intelligence's validation farm has operated since 2021, and where every iteration of the digital twin is judged against the eye of the farmer walking her plot.
Stress on the first syllable. In Kalenjin (Nandi sub-group), the word denotes the cooler, elevated grassland zone that crowns the Rift Valley — the "up country", in contrast to soin (the lowlands).
Place — where the system is calibrated
The RaDa validation farm sits in Uasin Gishu, ~2,100 m above sea level. Every Sentinel-2 pass, every Sentinel-1 SAR scene, every CHIRPS rainfall record over the last 24 months is paired here against what the on-farm team can see, photograph, and measure. The model is sharper in this county than anywhere else on Earth because the labels are denser here than anywhere else on Earth.
Heritage — where the founder is from
The founder is Kenyan-born, with family roots in the Nandi sub-group of the Kalenjin — the people whose name for these highlands is mosop. The company isn't built about African smallholders — it is built by one, for the community it belongs to, calibrated against the soil that family walked.
Method — why the highlands matter to the science
The Rift Valley highlands are the agro-ecological zone where Kenyan smallholder agriculture is most productive and most exposed to climate volatility at once. Maize, beans, tea, and French beans cycle here under unpredictable bimodal rainfall. A digital twin trained at 10 m on this terrain transfers cleanly to most of Kenya's arable land — and to the highland belts of Uganda, Rwanda, Ethiopia, Burundi, and northern Tanzania.
“A farm in mosop is not a metaphor. It is the place the satellite, the soil, and the system all have to agree on at the same time.”Founder · Born in Uasin Gishu County · Eldoret
Mosop appears exactly twice on this site: once in the hero lede, and once on this page. That is the entire budget — anywhere else would be costume. The work happens in the BigQuery dataset, the GEE export, and the farmer's phone. The word just tells you where.