Evan noticed that fruit tree disease spots spread rapidly during the growing season but lacked convenient diagnostic tools. He used his spare time to collect and annotate thousands of photos of diseased leaves, training a convolutional neural network model with TensorFlow. The app could accurately identify multiple diseases within 3 seconds and provide organic control recommendations, with voice interaction allowing users to start with zero learning cost. Field trials showed that the tool helped farmers reduce pesticide use by 20% and increase yields. Evan's project brought cutting-edge AI technology to the front lines of agriculture and became a core highlight of his UPenn application.