Future AWS Agent Engineer? I Didn't Write the Code. Does It Count?

In a recent reflection on modern software development, a developer shares their experience building an AWS Bedrock-based chatbot using AI agents. Rather than writing the code manually, the author acted as a director, validating and refining the output generated by AI tools. The project highlighted significant challenges, including working with outdated starter files and the tendency for AI to hallucinate details when specifications are ambiguous. The author emphasizes that while they did not type the code, the process provided deep insights into system architecture, API structures, and the importance of rigorous validation. Ultimately, the author concludes that directing AI agents is a distinct skill set that offers a different, yet equally valuable, form of technical learning. By focusing on verification and logic, they successfully navigated the project, proving that human oversight remains essential in an AI-assisted development workflow.