I didn't start with a grand plan. My first real program was a clunky Python script that scraped a website I cared about, and the moment it worked I was hooked on the feeling of making a computer do something useful.
Building tools for an academic-operations team taught me that the best engineering is invisible - it quietly removes hours of someone's week. Founding engineering at a startup taught me the opposite lesson too: sometimes the work has to be loud, fast, and ready for a traffic spike at midnight.
The throughline has always been the same: build the thing, ship it, watch real people use it, and let that feedback bend the next decision.
If you're early in your own path, my only advice is to keep shipping small things. The portfolio takes care of itself once the habit does.