All Projects
IN DEVELOPMENT
#MOBILE APP#MACHINE VISION#AGRI-TECH
AgroSense AI
GHANA · AgroSense Inc.
THE PROBLEM WE'RE SOLVING
Smallholder farmers across West Africa lose 30–50% of their harvests to crop diseases they can't identify fast enough. Agronomists are expensive and slow. Lab tests are out of reach entirely.
WHAT WE'RE BUILDING
We're building AgroSense AI — a mobile app that lets any farmer photograph a crop and get an instant disease diagnosis, on-device, without internet. The model runs on a $40 Android phone. The interface was designed for farmers who've never used a smartphone app before.
STACK
PyTorchFastAPIReact NativeAWS
On-device ML model
MobileNetV3 trained on 280,000 field images, quantised for offline inference on low-end Android hardware
14 disease classes
Built from a proprietary dataset assembled with 3 agricultural universities
Accessible UI
Icon-only navigation, audio feedback in Twi and Dioula, 3-step capture flow under 30 seconds
Cooperative rollout
Deployed via field agents trained to onboard farmers across Ghana and Côte d'Ivoire
PROJECTED OUTCOMES
94% detection accuracy targetedTarget
12,000+ farmers in rollout planTarget
40% crop loss reduction projectedTarget
WORK WITH USGOT A PROBLEM
GOT A PROBLEM
WORTH SOLVING?
Tell us about your project and goals. We're here to help turn great ideas into powerful digital experiences.