From Ledgers to Language Models

My career began in finance and data analysis, where I spent nearly three decades in the disciplines that demand precision: reconciliations, reporting, budget analysis, and the kind of scrutiny where a single misplaced decimal has real consequences.

Over 27 years, I worked across financial reporting, accounts management, and data analysis functions — building the kind of operational instinct that only comes from doing the work, not just studying it. That experience taught me how organizations actually run, not how they're described in a strategy deck.

As large language models matured, I recognized something most AI commentary was missing: the technology's biggest constraint wasn't capability — it was domain grounding. Generic AI tools produced generic, sometimes wrong, answers when applied to specialized fields.

I began applying my finance and data background to AI systems architecture — designing agents and research tools that were accurate because they were built by someone who understood the underlying domain, not just the underlying model.

Pulse AI became the umbrella for that work: corporate AI consulting for organizations, and a growing suite of focused AI products. The first of those products, Pulse AI for Physician, applies the same rigor-first philosophy to a field where accuracy is even less negotiable than finance — clinical medicine.

Today, Pulse AI sits at the intersection of two things I take seriously: deep domain expertise and genuine AI fluency. That combination is the throughline across every engagement, every agent, and every product I build.