I spoke with a software engineer whose company is being sold in 12 to 18 months.
Here’s why he’s building AI engineering skills now, while his job is still safe.
Call him Marcus. A few months ago he tried to build something with RAG, and in his words: “I realized I didn’t know what I was doing.” He put it down.
Five weeks of learning later, he had built a chatbot that queries a normal SQL database and a RAG database, with tool use holding the two together. It’s going to investors as a demo.
He called it “a new capability, a new superpower.”
But the build is not the interesting part. His situation is.
His employer is a B2B marketing company. Today the product is manual — customers drag and drop their landing pages by hand. The company wants to become AI-native, with a chatbot doing the work instead.
And then they want to sell, inside 12 to 18 months.
Marcus said it plainly: he has no idea where that leaves him.
Here’s what I told him. On the day the deal closes, there are two kinds of engineer in that building. The one who built the AI system the buyer is paying for. And everyone else.
The first one sets terms. “Happy to stay — here’s my price.” The second one waits to hear.
What gets him into the first group is not starting over. He already knows UI. He already knows CI/CD and infrastructure. AI engineering skills stack straight on top of that — chunking, context, tool use, agents, deployment. The blockers, not the buzzwords.
That’s also why the market pays what it pays. He sent me an AI engineering role at a well-known streaming company: $600k+ total comp. Those roles exist because very few people can actually build and run this stuff end to end.
Marcus started learning while he was comfortable and still employed. That’s the whole point.
Leverage is much easier to build before you need it.
Kirill
P.S. If your company is heading for an exit, that clock is your deadline, not theirs. AI Engineering Program: book a call here