I’m looking for a deep-learning specialist to build an end-to-end crop-disease detector that analyzes leaf photos and flags problems early. The model must recognise diseases on tomatoes, wheat, corn, potato, strawberry and blueberry, with the flexibility to add more crops later. Scope • Collect or curate a balanced image set (public datasets are fine) and perform robust preprocessing and augmentation in OpenCV or similar tools. • Design and train a Convolutional Neural Network in Python using TensorFlow / Keras. • Validate performance and tune until you reach strong, clearly reported metrics (confusion matrix, precision/recall and at least high-80s overall accuracy). • Package an inference script or small API that takes a leaf image and returns the predicted disease class along with confidence. • Supply clean, well-commented code, a brief README and any model weights so I can reproduce results on my machine. Nice-to-have If time permits, a simple Streamlit or Flask demo that lets users drag-and-drop an image and see the diagnosis would be great. I’ll provide quick feedback on each milestone, and I’m happy to discuss architecture choices—EfficientNet, ResNet or a custom CNN—whichever best balances speed and accuracy on our target crops.