Nowadays, 1 in 36 children is diagnosed with autism spectrum disorder (ASD) according to the Centers for Disease Control and Prevention (CDC) [52], which makes this condition one of the most prevalent neurodevelopmental disorders. For children on the autism spectrum who face substantial developmental delays, the trajectory of their cognitive growth can be markedly improved by interventions if the condition is identified early. Therefore, there is a critical need for more scalable screening and diagnostic tools, as well as the need to improve phenotyping to refine estimates of ASD symptoms in children. Here, we introduce AI4Autism: a 4-year project funded by the Swiss National Science Foundation, which aims to address the needs outlined above. In this project, we examine the potential of digital sensing to provide automated measures of the extended autism phenotype. This is accomplished using multimodal techniques based on computer vision and Internet of Things sensing, for the purpose of stratifying autism subtypes in ways that would allow for precision medicine. We present an overview of our main results so far, introducing datasets and annotations that we intend to make publicly available, as well as methods and algorithms for analyzing children's behaviors and producing an ASD diagnosis.