I build systems where technical, regulatory, and organizational constraints are tightly coupled — and failure comes from misunderstanding that boundary. That's not a philosophy I arrived at through theory. It's what I learned building inside SCIFs, shipping through FedRAMP audits, and scaling platforms that banks wouldn't tolerate going down.
Early on, I learned that the best engineering work is invisible — the system that never goes down,
the migration no one noticed, the security architecture that held under pressure no one anticipated.
That shaped how I think about what "done" means.
At Salesforce, I got to work on problems that didn't have precedent: building infrastructure automation
when the playbook didn't exist, then building inside a classified facility when the constraints were
unlike anything in commercial software. Both taught me that the discipline that gets you through
ambiguity is the same whether you're debugging a race condition or managing a government client crisis.
When I joined Blend, the challenge was different — not just build something new, but change what the company
was. That required rebuilding the team, the culture, and the product simultaneously, without losing the
enterprise clients watching closely. I found that I'm drawn to exactly that kind of work:
where the technical and the organizational have to move together, or neither moves at all.
Now at Abridge, the stakes feel the most consequential yet. AI that helps clinicians document care for veterans —
that's not an abstraction. Getting it right means better notes, less burnout, and more time at the bedside.
Getting it wrong has consequences. I think that's exactly the right kind of problem to be working on.
