Rey Columna
All guides

What a Hands-On Healthcare CTO Actually Does

I'm Rey Columna, CTO at Welby Health since April 2022. We do remote patient monitoring and care management for clinics. Hands-on, for me, means I run the technology org and still write production code. This is what that looks like.

By Rey Columna, CTOPublished 3 min read

On this page

What does a healthcare CTO own?

I set the product strategy and the technology roadmap, then run the change that follows. That covers AI, data, platform, infrastructure and security. DevOps, QA, R&D, frontend, backend and mobile all sit under one roadmap.

Healthcare adds a layer most CTO jobs don't have. Our software sits between nurses, patients, billing and five different EHRs. Each of those has its own rules, and none of them are ours to change.

What makes the job different in healthcare?

  • A wrong answer costs more. In clinical work, a confident wrong answer is worse than no answer. Anything that touches patient data has to show where its answer came from.
  • You work inside systems you don't own. Epic and athenahealth set the terms. We connect to them. We don't get to rewrite them.
  • Security and compliance are part of shipping. HIPAA and SOC 2 checks are part of how the team ships.

What does hands-on look like for me?

I build real parts of the product myself. Three examples.

Markus

Welby's clinical AI teammate for nurses. I designed and built it and wrote most of the code. All 171 frontend commits and 242 of the 303 backend commits are mine.

Nurses use it every day to draft outreach messages, flag likely device issues and condense long notes. It can reach 30+ clinical tools on request. It checks clinical values against the data it actually fetched, and when something is missing it reports that instead of guessing. More on the Markus page.

Agent Console

A desktop app for running a team of AI agents. I built it alone, and all 352 commits are mine. About a dozen people across engineering, billing, operations, data and IT use it every day. More on the Agent Console page.

An athenahealth Chrome extension

We shipped it in a few days. It tracks a behavioral health session and suggests the time-based billing code: 90832 for 30 minutes, 90834 for 45, 90837 for 60. It needs no login and collects no data. athenahealth approved it for their behavioral health sales demos.

How do you run a healthcare engineering team now?

We have seven engineers, and each one has a small staff of AI agents. One agent covers our Epic integration, one covers athena, one covers billing, and one does nothing but security and compliance.

The engineers decide what gets built and review what comes back. Every change still needs two human approvals before it ships, same as before. In healthcare, I don't want that bar to move.

When should a healthcare CTO step back from the code?

When you become the bottleneck. Or when coding turns into a way to avoid hiring, budgets and hard conversations, which are still the job. Your code also gets the same review and tests as everyone else's.

Leading the org comes first. I keep building because it's how I know what a week of work looks like, which estimates are soft, and what the tools can do today. I wrote more about that in Why I Still Ship as a CTO.

If you're working on the same problems, email me at hello@reyco.me.

Sources