Early identification of neurodevelopmental disorders in infants is crucial for timely intervention to improve long-term functional outcomes. Feasibility trials have established differences in toy interaction behaviors between healthy and at-risk infants. The proposed smart toy aims to improve quantitative measurements of these interactions to detect risks of developing motor delays. Utilizing a unique configuration of soft lossy force sensors, the toy shows promise in classifying toy interactions such as grasp, hit, squeeze, kick etc.