What happens to data engineering when AI can write the code?
In Season 4, Episode 2, host Aaron Phethean sits down with Julian to find out. Julian is a software engineer with deep roots in cloud computing and a sharp point of view on where his industry is heading.
His opening shot sets the tone. AI won’t replace data engineers, he says. Engineers who use AI will replace engineers who don’t.
From there the conversation moves fast. Where is AI actually delivering value inside Julian’s team? Why do some engineers quietly resist it? And what does that resistance reveal about identity, ego, and the way we define good work?
The episode tackles one of the most debated questions in modern analytics. Does AI really need a semantic layer? Julian’s answer is more interesting than yes or no. Today, probably. In two years, probably not. He extends the same logic to local AI models running on everyday laptops and predicts where the next wave of productivity gains will come from.
Then comes the sharper critique. Low code tools have made it so easy to build pipelines that business critical data work is now happening without version control, without tests, and with credentials baked straight into scripts. Julian wants engineering practices brought back into data, and the cost of ignoring this, he warns, is bigger than most teams realise.
The conversation closes on a prediction worth pausing over. Dashboards are dying. AI agents won’t just show what happened. They will explain why, and recommend what to do next.
If you lead a data team or build with data, this one is for you.
