This a mobile application that enables farmers to perform in-field diagnosis of crop diseases and also provide advisory services in real time. More information on the classification model used can be got here. The advisory part which is a chatbot, works with a model designed with state of art algorithms in natural language processing. Initially, this model was deployed on Google App Engine but due to the challenge of internet connectivity within the areas where smallholder farmers are, we re-designed this algorithm to run locally on the mobile devices without need of internet connectivity and still provide real time feedback. I led the implementation as the machine learning engineer while at Busitema AI lab with Dr. Godliver Owomugisha as the PI and collaborated with other people like Dan Turcza from Google.
- Python
- Computer vision
- Natural language processing
- Mobile App development
- AWS
This application was available on Google Play Store as AgroDFTool by the time of writing this documentation. For more information, please contact me at omarajonathan@gmail.com