Developing AutismCare was a deeply rewarding yet challenging journey. One of the primary hurdles we encountered was working with medical data, which required careful handling, standardization, and significant preprocessing effort. Since our project involved machine learning on brain scans, achieving accuracy without overfitting was difficult due to the limited amount of high-quality, labeled data. Platform constraints also added friction—working on cloud-based environments like Google Colab introduced limitations such as restricted public link sharing, runtime disconnections, and memory bottlenecks. Additionally, coordinating model training and deployment in such environments often led to technical interruptions. On top of that, maintaining momentum through long hours of debugging, learning new libraries, and juggling academic responsibilities was a mental and physical challenge. However, these setbacks taught us resilience, improved our collaboration, and ultimately helped us build a more thoughtful and impactful solution.