I build the data platforms healthcare companies run on.

Twenty years of it. Right now I own architecture, governance, and SOX compliance for the lakehouse behind a national health insurance marketplace. Before that I built the pharmacy data warehouse at a health plan serving more than ten million members.

The work has a shape that repeats no matter which industry I am in. Many sources send you the same thing, many different ways. A back office says one thing, a clickstream says another, a call log says a third, and all three are describing the same person. Most of the engineering is not in moving the data. It is in deciding what to believe, proving it, and being able to show your work afterwards.

That is what I have spent two decades getting good at: ingestion that survives contact with real partners, reconciliation you can put in front of an auditor, identity resolution across sources that were never designed to agree, and the governance that keeps the whole thing trustworthy while it grows.

What I believe about this work

A new source should be a configuration entry, not a new codebase. If onboarding your sixth partner costs what the first one did, the problem is architectural. Parameterise the ingestion, build reconciliation into every load, and the cost curve flattens.

Choose the design whose failures you can survive. When two approaches are close on merit, the tiebreaker is what happens when you get it wrong. A missed row-level policy denies access. A missed application check leaks data. That asymmetry should decide the architecture, not developer convenience.

Data nobody trusts is worse than no data. Reconciliation, lineage, and a dictionary somebody actually reads are not overhead on the real work. They are what makes the real work usable.

Stay close enough to build. I lead, review, and set standards. I also still write the code, tune the jobs, and get paged. The day I cannot do that is the day my architecture calls stop being credible.

Outside the day job

I build things. Homemade is a hyperlocal marketplace I designed, secured, deployed, and operate alone. It is deliberately small. It is where I keep my hands on production systems instead of only reviewing other people's.

I play chess at a level that keeps me honest, and I have never met a puzzle I could leave alone. Both are the same instinct as the day job: look at a position, work out what is actually true, and commit.

Get in touch

I am open to conversations about data platform leadership roles, particularly in healthcare. The fastest way to reach me is email.