AI-powered crop disease detection for Kenyan smallholder farms — Omdena Kenya Chapter (Product Owner)
# Early Detection of Crop Diseases in Kenyan Smallholder Farms 🌿
> **Led a 52-member Omdena team as Product Owner to build an AI-powered mobile application for real-time crop disease detection — achieving up to 100% classification accuracy across 4 crops and 20+ disease classes for Kenyan smallholder farmers.**
## Table of Contents
- My Role & Contributions
- Problem Statement
- Dataset
- Methodology
- Model Performance
- Deployment
- Technologies Used
- Team & Acknowledgments
- Links & Resources
- Author
## My Role & Contributions
As **Product Owner** for the Omdena Kenya Chapter, I led the end-to-end delivery of this AI-powered crop disease detection system. My responsibilities included:
- **Team Leadership** — Coordinated **52+ collaborators** across data collection, model development, and deployment task groups
- **Product Strategy** — Defined the project vision, success criteria, and delivery roadmap
- **Sprint Management** — Led sprint planning, daily standups, task prioritization, and retrospectives
- **Stakeholder Communication** — Served as the primary liaison between the technical team, Omdena leadership, and agricultural domain experts
- **Technical Oversight** — Oversaw the complete ML pipeline from raw image data collection through model evaluation and mobile app deployment
- **Model Selection** — Drove the decision to deploy YOLO based on its superior accuracy-speed tradeoff across all four crop categories
- **Quality Assurance** — Established evaluation benchmarks and ensured models met accuracy thresholds before deployment approval
## Problem Statement
Kenya's **7.5 million smallholder farmers** produce approximately **80% of the country's agricultural output**, yet they face a critical challenge: **limited access to crop disease diagnostic resources**. When diseases go undetected, farmers can experience **30–40% yield losses**, directly threatening food security and rural li …