Skip to main content

AI Engineering Cookbook

This section outlines best practices for using AI-assisted coding tools effectively. While AI may fall short in some development scenarios, it excels in many others. Getting the most from these tools requires more than just adopting the technology—it also means adapting your workflows to align with the strengths and limitations of AI. A pragmatic, open-minded approach is key to integrating AI into the software development lifecycle. Done right, this enables faster and more efficient product delivery. This section explores how to methodically incorporate AI into your workflow and highlights tools that can significantly improve outcomes.

📄️ AI-Driven Code Review with GitHub Copilot

I recently had a huge spike to get a large piece of work over the line. It was working a solo, 30 hour weekend, so whilst I had to ship unreviewed code, I also wanted a safety net to spot silly mistakes, or where I've just cut corners. GitHub's Copilot reviews is the ideal tool for this, and spot quite a few inconsistencies or bugs that could have been nasty to debug. Whilst I wouldn't recommend it as the only reviewer in a large team, it is invaluable as a first draft reviewer. With the various teams I'm working with, we all use it as a first step, address any issues, then share it for review with the wider team. Even in the last month, it's saved a lot of time and is great for upskilling.