Overview
Own the reliability of Harrison.ai's US teleradiology platform on GCP and Kubernetes, focusing on availability, failure modes, and fast recovery rather than scale.
What you'll do
- Own the SLOs for the teleradiology platform — define them with the people who depend on them, instrument them honestly, and use the error budget as a real input to what we ship and when.
- Run production on GCP and Kubernetes, and steadily make it duller: fewer surprises, smaller blast radii, faster recovery.
- Build the observability picture in Grafana so that when something goes wrong, the answer to "what's broken and who's affected" takes seconds rather than an archaeology session.
- Lead incident response, and write the postmortems that change the system rather than the person.
- Find the toil in the squad's week, measure it, and engineer it away.
- Make reliability legible to people who aren't engineers — radiologists, clinical operations, and the commercial team all need to understand what we're promising and whether we're delivering it.
- Work reliability into how we design and change things, rather than inspecting for it at the end.
What you'll need
- Real production ownership: responsible for a system that mattered when it broke, not just built it, but carried it, and stayed to make it better.
- GCP and Kubernetes operated under pressure, not only configured in a quiet moment.
- Grafana: build dashboards and alerts that people actually trust and act on.
- An SRE practice: SLOs with error budgets that really do gate releases, blameless postmortems, toil that gets measured and removed, capacity and change treated as engineering problems rather than ticket queues.
- The judgement to know where the Google SRE model doesn't apply to a small team inside a regulated medical device company.
- Genuinely excellent written communication, including handing incidents across a timezone gap and explaining reliability to clinicians.
Nice to have
- Experience running production in a regulated industry, where audit trails, change control and validated environments were part of the job.
- Healthcare or medical device experience, particularly touching ISO 13485, IEC 62304 or ISO 14971, and the ability to move quickly within a quality system.
- Experience on a team split across widely separated timezones.
- A background in clinical, diagnostic or other systems where downtime had consequences for someone's safety.
Details
- Remote within the United States.
- Covers US business hours; outside those hours the platform hands off to the Australian team. No overnight pager.
- Flexible hours and unlimited work-from-home, with some overlap with the Australian team to be agreed together.
- Distributed team across four continents; decisions and context are written down.
Read the full description and apply on the company’s own careers page.