Twelve years turning ad-hoc, legacy setups into modern, dependable cloud platforms; lately, bringing AI into how engineering teams work. The bridge between developers, operations, the business and management.
For twelve years I've been the person who makes sure software can actually be trusted in production. I started in backup and operations, moved through cloud and DevOps engineering across agencies, telco and product companies, and today I lead DevOps at Watts.
What I'm known for isn't a single tool; it's foresight and pragmatism. I'm good at reading where a product or platform is heading, spotting the blind spots early, and naming them clearly to a cross-functional team and to upper management. I don't sit on problems: I turn ideas into small, testable deliverables and prove them before we scale.
For the last three and a half years I've also served as a technical advisor to the Director of Software Engineering, shaping strategy, not just executing it. That's the level I want to keep building on.
"Asbjørn is energized, great at sparring and pragmatic when it comes to solutions. He knows what the 100% looks like, but accepts that 80% is sufficient at first, and the last 20% is a matter of planning." A colleague
Bridging developers, operations, the business and management. I translate between them, shape strategy alongside senior leadership, and keep everyone moving in the same direction.
Taking legacy, ad-hoc infrastructure to containerized, scalable Azure platforms (Docker, Azure Container Apps, infrastructure as code) that are cheaper to run and built to grow.
EU Cyber Resilience Act readiness, SBOM and vulnerability scanning, compliance-gated CI/CD, and the production-readiness judgment that keeps unsafe code out of production.
I've adopted AI seriously, and I think about it in three layers.
I co-led a programme with another tech lead to expand and upskill our developers in AI-assisted engineering, facilitating Claude Code licenses and the practices to actually use them well.
I built internal agents on Azure AI Foundry: a log-search and monitoring agent that calls my phone when something serious breaks, and a "DevOps junior" agent developers can simply message in Teams.
The highest-stakes layer, and one I believe should be governed and introduced with real care. With adversarial techniques like jailbreaks evolving fast, I help teams move toward it deliberately, with the guardrails and oversight that earn lasting customer trust.
It's the same instinct across all three: move fast where it's safe, and put the right guardrails where the stakes are higher.
A flagship product was ready to ship the week I joined, but on an outdated "build-a-zip, deploy-to-function-app" setup. I rebuilt it from the ground up as a containerized platform (Docker, Azure Container Apps), the company's first at the time. Scaling became far more efficient, compute cost dropped around 25%, and I moved every credential out of Git and the build pipeline to managed identity. It became the foundation that the wider backend, and every product launched since, now runs on.
I led our readiness for the EU Cyber Resilience Act. I introduced SBOM generation and vulnerability scanning, and built CI pipeline templates that automatically block non-compliant software from being released. Working hand-in-hand with QA and with sign-off from Legal, I wrote the CRA handbook our audit team now uses. Today around 150 pipelines run under these gates, spanning mobile apps, cloud backends, IoT devices, and embedded firmware.
At scale, our microservices were generating so much log data that logging cost 50% on top of every compute dollar. Instead of mandating cuts, I brought the data to the developers, who, it turned out, weren't attached to the logs, just to their ability to debug. We shipped a shared library with sensible production defaults (while keeping log levels easy to turn up for debugging), cut the near-useless noise, tuned sampling, and consolidated logging workspaces for volume discounts. The target was 30%; we reached nearly 50%, around $10,000 saved a year, and reinvested it into new AI initiatives. The bigger win: three product teams, DevOps and QA solving a pure engineering problem together.
A commercial IoT product was on track for production, but a readiness review I led showed it wouldn't scale beyond around 50 units, against a commercial target of 10,000 a year. I made the data-backed case to harden it first (security, code quality, an architecture built for volume) and aligned the team and leadership around getting it right. Catching that gap early turned a fragile prototype into a launch with a foundation that could actually grow.
I cook constantly, and when I travel I get to know a place through its food. Most days also include a walk with Skipper, my Cardigan Corgi.
Quietly exploring my next step in DevOps leadership or IT management.