From Excel to AI: Structure, Questions, and Becoming Who You Are

I didn’t start as a programmer. I started as someone people enjoyed asking questions to.

My entire life, people have pulled me aside to ask my opinion — about politics, about culture, about decisions that were sometimes far above my age. Teachers, friends, adults, colleagues. I was often the youngest person in the room, but somehow the one asked to make sense of it.

I didn’t know why. It didn't matter the topic — relationships, history, politics, economics, psychology. When I was about three, my dad sat me down and asked me to name the Axis and Allied powers of World War II. I answered him. Growing up, I always wondered what it was about my perspective that prompted adults to ask me adult-like questions.

Maybe some people are just wired to look for structure in the noise. Maybe some people are meant to organize what others feel but can’t articulate. Sometimes destiny doesn’t arrive as a lightning bolt — it shows up as a question.

Years later, I was a government major interning in the California Legislature. My job wasn’t glamorous. It wasn’t tech. It wasn’t strategy. It was Excel.

At the time, it felt administrative. Mechanical. Almost boring. But looking back, that spreadsheet was my first real experience with coding.

Because coding isn’t mysterious symbols on a black screen. Coding is structured thinking. It’s deciding what matters and what doesn’t. It’s creating consistent fields so chaos becomes searchable. It’s turning paragraphs into variables. It’s separating signal from noise.

When I highlighted the pros and cons of a bill, I was performing classification.

When I standardized language across hundreds of letters, I was normalizing data.

When I reduced emotional testimony into bullet-point arguments, I was extracting features.

I didn’t call it machine learning. I didn’t call it data architecture. I just called it doing my job. But in today’s AI-driven world, those same principles power intelligent systems.

AI does not thrive on chaos. It thrives on clean inputs. Clear categories. Repeatable logic. Consistent formatting. The same discipline I learned logging legislative mail is the discipline required to build reliable machine learning systems.

And maybe that’s why people asked me questions all those years. Not because I had all the answers — but because I instinctively tried to organize the problem before reacting to it. I look for categories. I look for definitions. I look for the columns before I fill in the rows.

So when I build software now — whether I’m structuring CSV pipelines, designing AI prompts, organizing metadata, or building tools that learn from information — I return to the same principle I learned in that legislative office:

If a human can’t clearly categorize it, a machine definitely can’t.

My coding philosophy didn’t begin in a computer science lab. It began with questions. It was refined in spreadsheets. And in a strange way, the more “boring” the data looks, the more powerful the system becomes.

Because underneath every intelligent machine is something very simple: a clean table, a clear structure, and someone willing to ask the right questions first.