SamS. Won

My work sits at the seam between rigor and pragmatism. For a decade I have built systems that have to survive contact with the real world — predictive models for global financial institutions, clinical-trial analytics for one of the world's largest biotech companies, and bespoke software for small businesses that quietly run on it every day.

I am, by training, a mathematician. By practice I am an engineer who prefers the unglamorous parts of the trade — the lineage of a dataset, the way an interface concedes to its user, the version-controlled history of a decision. I trust craftsmen over frameworks, and I think data-science work is better when it is conducted like a craft.

That conviction is exercised in two places, kept deliberately separate. By day, I am on staff at Amgen — embedded in the Operational Design Analytics team, where I turn historical and operational data into the foundation for planning the next generation of clinical trials. The work is consequential, slow, and properly regulated — the opposite of headline AI work, and the more meaningful for it.

Alongside, and intentionally separate, I keep a small consulting studio under the name Epsilon DS. It exists to give ambitious teams the kind of senior, hands-on data and software work that is otherwise difficult to find — bespoke machine-learning pipelines, point-of-sale platforms, internal tools, dashboards — each engagement conducted personally, not through layers of account managers.

Whether I am tuning a neural network, drafting a migration plan, or pulling an espresso shot, I bring the same posture: a quiet preference for precision over performance, and a tolerance for the long route when the long route is the honest one.