Autism spectrum disorder (ASD) can be defined as a neuro-developmental condition that affects communication, behavior and social interaction. Intervening early in the lives of children with ASD has the potential to enhance their cognitive abilities and alleviate autistic symptoms. Numerous clinical studies have indicated the presence of distinctive facial features in children with ASD, often manifested through alterations in facial anthropometry. This study introduces a pragmatic screening approach for ASD using facial images, employing deep learning techniques on a publicly available dataset consisting of face images of diagnosed children with ASD.