Shashank Madala

I research speech-based measurement of autism symptom severity at the Yale Child Study Center, working with a longitudinal clinical corpus of 197 children. My primary project addresses a measurement problem rather than an accuracy one: existing severity models treat a child's set of recordings as an ordered sequence, so the same child can receive scores up to 10 points apart depending only on the order the recordings arrive in. I designed an order-invariant architecture that eliminates this artifact by construction, using roughly 3,000 parameters against the original's 1.2 million while matching its accuracy and returning one stable score per child. The work is first-authored and under review at an IEEE undergraduate research conference.

Alongside this I contribute to the lab's parent–child dyadic synchrony work, analyzing acoustic and behavioral alignment during recorded interactions, and I support preprocessing and annotation workflows on the longitudinal interaction data.