The new AI engineer's toolkit

The role of the staff and principal engineer is fundamentally changing. It is no longer just about writing code from scratch. Now, it is about being an AI orchestrator.

I have spent the last few months deeply crafting and refining prompts to supercharge my team’s productivity. We are moving from typing lines of code to crafting powerful prompts, verifying AI-generated solutions, and integrating tools like GitHub Copilot into our daily workflow.

The real skill shift

The real skill is shifting from “how do I write this function?” to “how can I architect this system to leverage AI for maximum efficiency and innovation?”

This process has become so central to our work that we treat our prompts as a first-class citizen in our codebase.

Practical starting points

Save and share your best prompts. We use .prompt.md files to store reusable prompts for common tasks. This is a game-changer for consistency and onboarding. When a new engineer joins, they immediately have access to the team’s accumulated prompt expertise rather than starting from scratch.

Treat prompt files like code. They go through review, they get versioned, they get improved over time. A well-crafted prompt that encodes your architecture patterns, naming conventions, and error handling approach saves hours every time it is used.

Build a prompt library, not just individual prompts. Slash commands in VS Code using .prompt.md files let you create IDE-level shortcuts for common workflows: generating unit tests, adding API endpoints, adding OAuth providers. This encodes your best practices directly into the AI workflow.

The engineers who figure out how to systematically leverage AI (not just use it ad hoc) are becoming dramatically more productive. The toolkit is here. The question is whether you are treating it as a toy or as infrastructure.