← All posts

What unexpectedly carried over from my past

When you switch into something as different as ML, the first question - the one other people ask and the one you ask yourself - sounds something like this: why did any of your previous experience matter at all?

I know people who try to distance themselves from their past when they enter a technical field. As if it only gets in the way. Or as if it is somehow embarrassing not to have come through the classic math-and-CS route.

For me, it turned out differently.

The mathematics I already knew

My first degree was in personality psychology. It sounds far away from machine learning. But there is something many people do not realize: mathematics is a core discipline in psychology. It is exactly what turned it into an independent science.

Higher mathematics, mathematical methods, analysis of variance and statistical data processing were not optional extras. They were part of the foundation. Course papers and final theses were defended with full mathematical treatment of data. It was living practice, not an abstract class.

So when I met probability, distributions and hypothesis testing again in ML, they were not unfamiliar. The foundation was already there. It had simply been buried under the years and had started to look irrelevant.

It turned out to be very relevant.

Working with people teaches you to ask why

Several years in sales and team management are, first of all, years of working with needs. You learn to hear not only what a person says, but what they actually need. You learn to think not “what can we offer” but “what will solve their problem.”

In ML, this is called problem framing. And it is one of the most underrated skills in the profession.

A technically strong person can build an excellent model for the wrong problem. I have seen that happen. When you know how to figure out “why this is needed at all” first, the system architecture becomes clearer even before the first line of code.

Explaining is also an engineering skill

There is a common misconception that if you can explain something complex simply, you probably do not understand it deeply enough. In reality, the opposite is true.

Working with different people - from clients to teammates - taught me to stay focused on the essence and not hide behind terminology. Technical projects also need to be explained clearly, especially if you want people to actually use them.

Past experience does not reset to zero

If you are changing professions and you come from a “non-technical” background, there is a good chance you already know something that many people who came directly through CS are missing.

That does not mean hard skills are optional. They are not - and the gap in computer science is real. It has to be closed methodically and without illusions.

But previous experience does not disappear. It simply waits for the moment when it becomes useful.

I have already had several of those moments. And I suspect there will be many more.