Teen builds app to help families read autistic cues
Shashank Madala built a free application in New Jersey that helps caregivers interpret the emotional signals of autistic children.
Madala watched adults frequently misunderstand his autistic cousin during family gatherings because they guessed wrong about his needs. The boy eventually stopped trying to communicate because he felt ignored by those who loved him the most.
He decided to teach himself machine learning at age fifteen to create a better solution for these difficult family situations. His early versions provided single answers that he realized repeated the mistakes his family had made with his cousin.
The rebuilt application now offers multiple possible readings while surfacing areas where the software is least certain about feelings. This design slows adults down and prevents them from accepting one quick answer without considering other possibilities.
Madala programmed the tool to flag uncertainty and leave the final decision to the person who knows the child best. A person holding a phone camera near the child receives suggestions for how to respond appropriately to emotions.
“The most useful thing I have ever built came out of paying close attention to one person”

