I really relate to this. I built my HintWordly website using Cursor AI—even though I’m a DevOps engineer and comfortable with AWS architecture and deployment. Getting the system live was smooth, but now I’m hitting a wall with code optimization. Since most of the code was AI-generated, I didn’t build the logic layer by layer myself. Now I struggle with things like reducing redundant API calls, improving query efficiency, and refactoring components to avoid unnecessary re-renders. Even tracing performance bottlenecks (like high response time or memory usage) takes more effort because I’m not fully familiar with every part of the codebase. like in that way? It really shows that while AI helps you build fast, optimizing and scaling the code still requires deep understanding of how everything works underneath.