I spent two years studying electrical engineering before I admitted to myself that I wanted to write software.
By the third year I had passed the hard exams and I knew where the degree led. My place was state subsidized, so it cost me nothing to stay. Everything about my situation argued for finishing. The only argument on the other side was that I kept spending my free time programming, and I kept feeling more alive doing that than anything on my syllabus.
So I left. Not by drifting out, but by moving to a different university, in a different field, on a paid contract. I went from studying for free to paying for my own education, and I covered it by working extra hours alongside my classes.
That is the part these stories usually skip. Leaving was one decision followed by years of small ones. Shifts that ran late before an early lecture. Weekends that were not weekends. Tuition coming out of money I had earned rather than out of nowhere, which has a way of making you take attendance seriously. Passion got me to the decision. It did not get me through what came after.
What I feared most was that the two years would simply vanish. They did not.
Electrical engineering does not teach you to write code. It teaches you how to reason about a system you cannot see directly. A circuit does not print an error message. When it misbehaves, you form a hypothesis, you measure, you narrow the space, and you keep going until the behavior makes sense. Nobody hands you a line number.
Years later, that is exactly what a production incident feels like. Something is wrong, the logs are incomplete, several systems are involved, and the only way through is disciplined reasoning from evidence. I had been practicing that long before I had a job title that rewarded it.
The same is true of constraints. In hardware, nothing is free. Every choice costs power, heat, space, or time, and you feel it physically. Software lets you pretend otherwise for a while. Coming from the other side, I never fully believed the pretense. Latency is real. Memory is real. A network call is not magic.
Paying my own way taught me something too. When your time is scarce, you stop confusing motion with progress. You learn to find the thing that actually matters in a piece of work and do that first, because there is no spare evening to absorb a mistake.
None of this was on a transcript. All of it transferred.
If you are two years into a program and something is not sitting right, three things are worth holding onto.
The years are not deleted. Whatever you have been studying has been shaping how you think, and that does not reverse when you change buildings. You are not starting from zero. You are starting from somewhere unusual, which is often better.
Switch on evidence, not on a bad week. My clearest signal was not that I disliked my classes. It was that I was already doing the other thing without being asked, in hours nobody was grading.
Count the real cost first. Know what you are giving up, in money and in time, and decide whether you are willing to carry it. A decision you have priced out survives the hard years. One made on enthusiasm alone usually does not.
I expected to spend my career quietly hoping nobody asked about the degree I did not finish. Instead it turned out to be one of the more useful things I own.
Leaving a field is not the same as losing it. I stopped studying electrical engineering. I never stopped thinking like someone who did.
