Are you a data engineer watching AI do your job in minutes?
Here’s why one engineer I spoke with is moving into AI engineering now — not in three years.
I’ll call him Sameer. Data engineer at a large insurance company, 3+ years of building production pipelines. Python, SQL, Databricks, AWS, Snowflake.
Six months ago, his team was pasting SQL queries into a chat window to speed things up.
Today, an AI agent with access to their data platforms refreshes an entire dashboard from one prompt.
His words: “Claude is doing the work. It’s literally doing my work in minutes.”
And the question he’s asking is the one most data engineers are avoiding: why would companies need this many data engineers?
So I played devil’s advocate on our call.
If AI is automating data engineering, what makes him think AI engineering won’t be automated next?
He didn’t flinch: “It very much could.” An engineer has to be forward-thinking, he said. When the field shifts again, you shift with it.
Then he said the line I keep thinking about:
“I’m not thinking about what’s happening tomorrow. I’m thinking about what’s happening three years from now.”
Here’s what I told him.
Data engineering is one of the best starting points for AI engineering.
RAG — the technique behind most enterprise AI systems — is about retrieving the right data and getting it to the model. That is data work.
Vector databases sound exotic. They are another way of storing and querying data. A data engineer can become the vector database expert faster than anyone else in the room.
The pipelines, the orchestration, the data quality instincts — all of it carries over. AI engineering skills stack on top.
Watching AI do your daily work is a signal. Sameer chose to read it three years early.
Kirill
P.S. Feeling AI put pressure on your career? See if you qualify for our AI Engineering Program: https://aiengineeringprogram.com/