Hot take: Don't become a data scientist. Become a data engineer.

Hot take: Don't become a data scientist. Become a data engineer.

Early in my career, I was an oil & gas field engineer.

One of my mentors told me he chose production engineering for a simple reason: production engineers are the last to get laid off.

Not the most glamorous role. Lots of field time. No cushy office.

Meanwhile, there's an old joke about drilling engineers:

What's the difference between a drilling engineer and God?
God doesn't think he's a drilling engineer.

Drilling engineers were the rockstars. High prestige. Big comp. Until oil prices tanked — and they were first out the door.

Here's what that taught me about data:

When budgets get cut, the data keeps flowing.

Someone has to maintain it. Someone has to keep the pipes flowing. That's the data engineer.

Think about it like your house:

Your heated pool heater breaks? $10k repair. You might wait months. Maybe forever. (Sound like any dashboards you know?)

But sewage backs up on your floor?

You're calling a plumber right now. Price is irrelevant.

Data pipelines can be like utilities. Nobody thinks about them — until they break.

Data engineers build job security that survives any market.

Be the plumber.

And here's a twist with AI. When you look at the three traditional data archetypes - data visualization, statistical modeling (aka data science) and data engineering:

Data engineers are the most equipped to actually deliver results with AI.

Not data scientists. Not data viz specialists. Not prompt engineers.

Data engineers.

Why? Because real AI value isn't a notebook. It's not a dashboard. It's a production grade system.

For example:

An autonomous accounts receivable agent that chases down overdue invoices, prioritizes outreach by churn risk, and routes escalations to the right rep — pulling live data from your CRM, your billing platform, and your ERP — doesn't just need a model. It needs pipelines, orchestration, API integrations, error handling, and something that won't silently break at 2am.

That's an infrastructure problem — and among the three archetypes of the data world, data engineers know infrastructure best.

Data engineers aren't just keeping the lights on anymore.

They're building the flashy robots.

Data engineers are getting to have their cake and eat it too. Defensively, they're the indispensable plumbers keeping everything flowing. But now? They're also front and center on stage — the ones actually shipping AI that works in the real world.

Photo: one of the sunsets from the rig where I worked in Argentina.