A Hackaday article reports that [Pmalfa31] applied an ESP32 and machine learning to grading tomatoes. The system knows in advance whether standard or cherry tomatoes are being processed and uses two different sets of learned data depending on the selection. The program receives raw data from an optical sensor, then processes it to remove empty belt images, compute statistical information, and group readings for a single fruit. The program includes heuristic checking of validity: it knows the approximate tomato size, so the code notes if a fruit seems too big for a cherry tomato or too small for a standard tomato, which reduces but does not eliminate miscategorizations. The code is a stack hog, so the program must request a much larger stack. A web-based simulator is available for trying the system without hardware or tomatoes.