A seven-year software engineer shares how AI tools have fundamentally changed his development approach since October 2025. After years of emphasizing clean code and architectural patterns, he discovered that Claude Code and Cursor enabled him to focus on problem-solving rather than writing code manually. His journey began with skepticism about tools like GitHub Copilot and Cursor, but he eventually embraced AI for both personal and professional projects.

His personal projects showcase remarkable productivity gains: migrating from Jenkins to GitHub Actions, consolidating multiple repositories into a monorepo, and switching from Dagger to BuildKite with Bazel. He's built everything from Discord bots to React Native apps, including a Claude Code multiplexer and tools for expense classification. At work, he's automated operations, created bespoke analysis tools, and rapidly implemented features that previously required extensive manual coding.

The shift required rethinking his engineering philosophy. He realized that for most projects, especially those he's paid to deliver, the primary goal is business value rather than code aesthetics. AI handles the implementation details while he focuses on design and review. Testing and documentation have become the new bottlenecks, as he must verify AI-generated outputs. He believes the industry needs better tools for testing and documentation to fully realize AI's potential in software development.