Where it started
I didn’t plan to end up automating DNS records and certificate renewals. My background is in molecular biology and histopathology—processing FFPE specimens, staining sections for diagnostic use, purifying proteins from pollen extracts for a bachelor thesis. None of that looks like infrastructure work on paper.
But looking back, the thread was always there.
The pattern I kept running into
At Dr. Ulm GmbH, part of my job was histological processing—grossing, embedding, and staining specimens for diagnostic use. Alongside that, I worked with local IT on digitalising lab workflows that had run on paper and manual handoffs for years. It was my first real exposure to the idea that a workflow doesn’t have to stay the way it’s always been just because it’s always been that way.
Later, as a Product Specialist at Sakura Finetek, I supported the Tissue-Tek Genie immunohistochemistry autostainer—first-level support, SOP creation, and a lot of root-cause analysis under time pressure with a customer waiting on the other end. That job taught me a specific kind of discipline: don’t guess, isolate the variable, document what you found so the next person doesn’t have to start from zero.
During my Master’s in Medical Biology, the same instinct showed up again in the lab. Instead of relying on closed-source tools for data evaluation, I automated data capture and visualisation with Python (matplotlib, pandas) and R (ggplot2). It wasn’t framed as “automation” at the time—it was just the fastest way to stop repeating manual analysis steps by hand.
The actual transition
In 2024, I moved into a Technical Assistant role at Paris Lodron University Salzburg, working across the Informatics and AI & Human Interfaces departments. The job description could have been written for the pattern I’d already been living: replace manual, outdated workflows with automated, auditable ones.
DNS records maintained by hand in a UI, certificates renewed (or forgotten) manually, secrets scattered across configuration files with no rotation—these aren’t so different from a lab process that depends on someone remembering to do the right thing at the right time. The fix is the same shape too: make the correct behaviour the automatic one, and leave a trail so mistakes are visible instead of silent.
What carried over, and what didn’t
What carried over: structured troubleshooting, the discipline of documenting why a decision was made and not just what was decided, and a genuine comfort with translating between people who don’t share a technical vocabulary—lab staff, IT departments, customers, whoever’s in the room.
What didn’t carry over automatically: the technical stack. Ansible, FreeIPA, ACME, OpenBao—all of that had to be learned from a standing start, on the job, the same way I once had to learn ion exchange chromatography from a standing start. Turns out that’s a transferable skill too.