Autism Spectrum Disorder (ASD), commonly referred to as "autism," is a psychiatric condition that affects a person's linguistic, cognitive, and social abilities. It's a prevalent disorder, with approximately 1 in every 54 births being diagnosed with ASD, and about 1% of the global population living with it. Unfortunately, despite its prevalence, the cause and cure for autism remain unknown, posing significant challenges to parents who suspect their child might have ASD. Early diagnosis of autism is crucial for a child's development, but it can be incredibly tough since symptoms manifest as the child grows. Typically, diagnostic tests conducted on children between the ages of 2 to 3 years are less reliable than those performed on children aged 4 to 5 years. This creates a worrying situation because early diagnosis is vital for autistic individuals to reach their developmental milestones successfully. Autism is often characterized by difficulties in social interaction and communication, making it challenging to diagnose accurately even with advanced tools like the ADOS and ADI. This work addresses the concerns surrounding autism diagnosis by focusing on improving the diagnostic pipeline. It involves training and testing machine learning models i.e., Random Forest with Standard scaler using an autism spectrum disorder dataset to identify the most significant indicators of autism in toddlers. The goal is to develop a quantitative approach to aid in early screening and subsequent treatment, as timely intervention can help mitigate long-term symptoms associated with autism. By leveraging machine learning, this work aims to provide valuable insights into diagnosing autism effectively and facilitating better support for individuals with ASD and their families.