AI Engineer is Just a Data Engineer Who Builds AI Apps - Change My Mind

To Nha Notes | Oct. 20, 2025, 9:34 a.m.

🚫 Stop Looking for AI Developers or Data Scientists to Create AI Systems!

I can't repeat this enough: 80–90% of GenAI work is already being done (or can be done) by Data Engineers.

Think about it...

What Really Happens When You Build a RAG Application?

When you build something like a RAG application, where's the AI magic, really? Let's break down what you actually do:

➡️ Ingest data (PDFs, APIs, etc.)
➡️ Extract and transform text into JSON
➡️ Create embeddings
➡️ Store them in a vector DB
➡️ Build an app to query it

All of this is engineering. Setting up pipelines, handling APIs, managing data flow, optimizing embeddings and retrieval. That's what engineers do.

So Where's the "AI Part"?

The "AI part"? That's just using a pre-trained model (like Mistral or Llama). You don't need to train it. The math to understand it all is quite simple. You don't need a PhD. You just use it.

The Reality of GenAI Work

For 80–90% of GenAI use cases:

• RAG applications
• AI Agents
• Chat over documents
• Semantic search
• Automation

👉 It's all a data engineer's job. No special AI wizard skills needed.

OK, we might call this engineer "AI Engineer" now. I can live with that.

What About the Remaining 10–20%?

We already have pre-built, downloadable models that handle that just fine.

It's not about "AI magic".
It's about knowing your data and building solid systems.

That's why I always say: Everything is Data Engineering.

Want to Build It Yourself?

For all engineers who want to understand what's actually happening under the AI hood, and build it themselves: