For our AWS-sponsored HCI capstone, my team researched how novice “vibe-coders” move from AI-generated prototypes to cloud-backed applications. Through interviews and contextual inquiry with 8 participants, we found that users relied heavily on AI for technical help but struggled with AWS terminology, service selection, cost uncertainty, and backend setup. Our research reframed the problem from tool integration to a larger onboarding gap: users understand what they want their app to do, but not how that translates into cloud infrastructure. This project is continuing next quarter with prototyping and usability testing.