← Blogs

What finance taught me about building data systems

A career shift from financial reporting to engineering shaped how I think about trustworthy data.

I began my career in finance, working on accounting, reporting, and the processes that turn raw records into numbers people use. Moving into engineering changed the tools I use, but it did not change the central question: can someone understand and trust the result?

A financial statement is not just a number. Behind it are source records, transformations, judgments, checks, and a point in time. If a result looks wrong, someone needs to trace it back and explain what happened. That instinct has stayed with me as I have worked on data pipelines and platforms.

I think about reliable data systems in three parts. First, make the path from source to output understandable. Second, treat corrections and changing definitions as normal events rather than exceptions. Third, make it possible for the people using the data to see where uncertainty or missing context affects the result.

The same ideas matter when an AI system retrieves information. A relevant document is only useful if it is appropriate for the question and its source can be understood. Good infrastructure gives people—and the systems they build—a way to inspect the evidence.

My transition from finance to engineering was not a switch from business thinking to technical thinking. It gave me a way to connect the two.