Client story · Hospital medical physics team · January 2026
Shipping a Hospital's Imaging App, with versioned containers.
A hospital medical physics team's X-ray image analysis app moved from manual deployments to container images built automatically for every release, ready to share with other hospitals.
The situation
A team of researchers had built a Streamlit app for X-ray image analysis. The app worked; getting it into production didn't.
Deployments were manual, and there was no reliable way to tell which version was running where. That made every update and every fix a risk.
Client
Hospital medical physics team
Stack
What we did
- 01
Name the bottleneck
What's working, what isn't, what's next. The problem wasn't the analysis code: it was how code reached production.
- 02
Images per release
A GitHub Actions workflow builds a versioned container image for every release and pushes it to GitHub's container registry.
- 03
A first release, end to end
A test run through the whole pipeline to v1.0, so the team saw the process work before relying on it.
- 04
Shareable with other hospitals
Published images, so another hospital can pull a specific version and run it without building anything.
- 05
Two compose files
Development builds the image locally; production pulls a pinned version. Both with health checks.
- 06
What's next
Recommended next step: automated HTTPS certificates for their internal domain.
The result
- —Every release is a versioned, reproducible image, from v1.0 onwards.
- —Deploying means pulling a tag; rolling back means pulling the previous one.
- —Other hospitals can run exactly the same version without building it.
- —Continuous integration removed the team's main operational bottleneck.
Want results like this?
A 30-minute decision call turns your situation into a plan.