Potato Disease Classification · Computer Vision & Deployment

Inspired by PlantVillage and industry tutorials, I implemented a full ML workflow for potato disease detection: from preprocessing and CNN model training to FastAPI serving, React/React Native interfaces, and deployment on GCP with TensorFlow Serving & TF Lite.

Highlights

• Data preprocessing and training pipelines in Jupyter Notebooks.
• Custom CNN model achieving robust leaf disease classification.
• FastAPI backend for inference with Dockerized deployment.
• Frontend in React (web) and React Native (mobile).
• TF Lite model deployed on GCP Functions for mobile efficiency.
• Explored scalable hosting using TensorFlow Serving + Docker.

Stack