Imagine an integration analyst at a large cancer hospital, a composite character rather than a real colleague. Call her Asha. She joined the IT department to connect systems: registration to billing, billing to the patient portal, the portal to appointment reminders. Within three years most of her connections ran as small cloud functions that woke up on events. Nobody gave her a title change. Yet her working life had quietly become what some engineers call ServerlessOps, and the skills it demanded were different from the ones she was hired for.
This piece follows her first year as a way to map the skills a healthcare technologist can build. Everything here happens in the back office, far from the wards, and none of it is clinical. The point is to show how a person’s skills grow when the plumbing underneath a hospital changes shape.
Month one: learning to see the invisible plumbing
Asha’s first discovery was that no one could say how many functions existed. Some had been created by a vendor, some by a contractor who had left. She started a spreadsheet: what wakes this function, what does it change, who owns it. That inventory was dull and enormously valuable.
The Ops Atlas describes the same opening move in its ServerlessOps write-up: spend the first stretch of a programme cataloguing triggers, failure queues and timeout settings before touching anything. If you want the full framing from the source, read the Atlas page on operating serverless workloads. Skill number one, then, is patient cataloguing. It looks unglamorous on a CV and shows up in every incident review.
Month three: the day a message arrived twice
The first real scare came when several patients received the same appointment confirmation two times. Nothing clinical was affected, but the call centre lit up. Asha traced it to a function that timed out after sending the message and before logging that it had. The platform did what platforms do and tried again.
What she learned that week is the second skill: designing for repeat. She worked out which real-world fact should identify one confirmation, the booking itself, and stored a note against it before sending. She also learned why identifying by the platform’s own message number would not have saved her, since a manual replay hands out new numbers. Few people teach this in a degree. It is the kind of lesson that makes colleagues trust you.
Month five: understanding what the failure queue is really for
Asha found a holding queue attached to the reminder function. It had been collecting failed items for months and nobody opened it. Her fix was almost embarrassingly simple: name an owner, add an alarm for the age of the oldest entry, and keep items for longer than the source retains them, so there was always something to examine after a long weekend.
That small habit became her third skill, owning the unglamorous leftovers. It resembles the way an oncology unit’s coordinators track every referral until someone has closed it. Readers curious about that coordination mindset in a care setting may enjoy our look at how a personalised treatment plan is assembled across teams; the systems analogy is loose, but the discipline of keeping every handoff visible is the same.
Month eight: money, and the dependency that slowed down
When a downstream records service became sluggish, Asha’s functions did not fail; they waited, and because billing follows running time, the monthly invoice jumped. She learned to read a bill as an operational signal. A slow dependency can raise cost sharply and push many instances running at once toward the account’s ceiling.
So her fourth skill was cost literacy. She began tracking cost per million events alongside reliability numbers and put a monthly slot in her calendar to look at both with finance. The Atlas calls that rhythm of review the fourth rung on its maturity ladder, between knowing your system is safe and tuning it with evidence.
Month eleven: rehearsal
By the end of the year Asha ran a small exercise she called a duplicate storm. She replayed a batch of test messages and confirmed that nothing happened twice. It took an afternoon, it embarrassed one function, and it gave the hospital something a policy document never could: a demonstrated result.
The Atlas treats such a rehearsal as the threshold for its third maturity level for critical paths. You can adopt the idea without the vocabulary: if a function touches money or appointments, test the repeat case on purpose and write down the outcome.
The skills, collected
- Cataloguing: knowing what exists and who owns it.
- Repeat-proof design: picking a business fact to identify each action.
- Ownership of failure paths: alarms, retention and a person who looks.
- Cost literacy: reading invoices as evidence about behaviour.
- Rehearsal: proving the safe behaviour before it matters.
None of these needs a certification, and the Atlas is clear that none exists for the discipline. What employers can see is a portfolio of small, written, tested decisions.
Where to start if you work in a hospital
You do not need Asha’s job title. Offer to inventory one workflow that is close to patients’ day-to-day experience, such as appointment messaging or portal notifications. Write down the triggers, the failure paths and the owners. Share it with your manager and with whoever handles vendor contracts.
Patients rarely thank anyone for a message that arrived once. They notice, painfully, when it arrives twice or not at all. If you are weighing where a hospital deserves trust, our note on why accreditation matters when evaluating hospitals overseas makes a similar point from the patient’s side: the reliable basics behind the scenes are what lets people focus on care.
Asha’s story is illustrative rather than reported, and this article offers general career orientation, not advice about any medical or legal matter.
