Professional

I have worked in IT for over 25 years. While studying, I built computers, taught Unix and networking courses at the TU Ilmenau computing centre and delivered my first projects. Development and systems administration followed; since 2018, my focus has been project management, Scrum and testing. I connect business stakeholders, development and operations, helping projects make a smooth transition into everyday use.

Project management

Proposals, planning, budgets and delivery records. I clarify requirements and turn them into work packages a team can deliver. I coordinate with clients, developers, operations and support through acceptance and handover. Agile where it fits; traditional approaches where they are needed. Certified at IPMA Level D (GPM (external website, opens in a new tab)).

Scrum Master

I facilitate daily stand-ups, refinement, planning, reviews and retrospectives, including for two teams at once. I plan capacity and leave, remove obstacles, support releases through production approval and coordinate with product owners and business stakeholders, including during PI Planning (external website, opens in a new tab). I help new colleagues settle in and act as a mentor. I take retrospectives seriously: we record what we learn and put it into practice.

Test management

Test design based on business specifications, regression, acceptance and user acceptance testing. Test organisation in Jira and Xray (external website, opens in a new tab), test data analysis, risk-informed release decisions and test documentation. Tools include Xray, Playwright (external website, opens in a new tab), Postman, SoapUI and PowerShell. If a tool is missing, I build it. A recent example is a status toolthat evaluates test accounts and test data in real time.

Technology

I understand what happens beneath the surface: software architecture, Unix and Windows, networks, databases, directory services, virtualisation, cryptography and PKI. My experience spans security systems, critical infrastructure through the smart meter rollout in Austria, and insurance IT with portals for customers, partners and municipal clients.

From 2027, I am available for project management, Scrum Master roles, test management, PMO and methods, project and start-up support, and consulting with a clear remit. AI is welcome in the toolbox. I am open to freelance work and permanent employment, based in Thuringia or working remotely. Get in touch.

2021 – 2026

Inverso GmbH, Ilmenau, Jena and Munich
Project manager and business consultant for portal and mailbox solutions at Versicherungskammer Bayern: proposals, billing and client coordination; Scrum Master for two teams; most recently, customer portal test management, release approvals and purpose-built testing tools

Case study: AI in test management at Inverso →

Inverso

2018 – 2020

Cuculus GmbH, Ilmenau
Professional Services Project Manager, delivering cryptographic solutions for critical infrastructure. In Austria’s smart meter rollout, Cuculus provided ZONOS as the central platform connecting meters, public key infrastructure (PKI) and data analysis. I managed this part on behalf of the main contractor, Ericsson, working with everyone involved to develop solutions for data collection, queries and key management.

Reference (German)

Cuculus · ZONOS

2003 – 2018

Saab Medav Technologies GmbH (formerly TeWiSoft and MEDAV), Ilmenau and Nuremberg
Development engineer, application engineer and systems administrator; complex security systems for data acquisition and analysis

Reference (German)

Saab

2001 – 2003

prokent AG, Ilmenau
Development engineer; automation of acceptance and return processes. The business later became part of what is now Diebold Nixdorf.

Nixdorf

1992 – 1999

Alongside my studies
MEGware GmbH: computer assembly, maintenance and customer support. TU Ilmenau computing centre: teaching Unix, internet, networking and TeX courses. Ilmenau/Suhl adult education centre: introductory computing courses for international students. Independent IT projects, including a payment reminder system for a client in Suhl

Computing centre

1991 – 2002

TU Ilmenau
Diplom in Electrical and Information Engineering, specialising in biomedical engineering and medical informatics

TU Ilmenau

1989 – 1991

Apprenticeship as an electrician
Kali und Bergbau Roßleben

The contact page has my CV and certificates as PDFs.

I have worked with AI every day for four years, using OpenAI, Anthropic and Google alongside one another. My everyday agent tools are Codex, Claude Code and Antigravity. I choose the right provider for each task.

What AI does well, and where it fails

AI can write, organise and check in minutes what once took hours. Its failures follow familiar patterns: reconstructing plausible answers instead of consulting the source, or treating a tool’s response as proof. My rule is simple: checking against the source is part of completing the work.

How I keep it under control

Clear guardrails. Each agent has explicit limits on what it may read, change or never do. In the testing workflow described here, AI works in a separate environment without direct access to sensitive systems. Only reviewed tools are deployed there, after approval. External actions remain with a person: emails, commitments and deletions. Changes are backed up, checked and documented. Mistakes still happen; restricted access, backups and independent checks are intended to limit their impact and catch them early.

Case study: AI in test management at Inverso

Regression testing for the customer portal was initially carried out manually in Xray . I built an AI-assisted workflow to compare test cases automatically with requirements in Jira and interface designs in Figma , raising Jira tickets where clarification was needed. Consistent test data across interfaces and maintained test accounts for each scenario were the main challenge. I built AI-assisted tools to handle those tasks too. Regression testing then ran automatically every four to six weeks, producing a report and draft tickets for deviations.

In this workflow, AI never had direct access to sensitive systems. An AI-assisted testing environment generates tools and test cases from abstracted evidence rather than real customer data. After review and staged approval, those tools run in the sensitive environment with read-only access and only when authorised. The team benefits from AI’s speed without sending customer data to an AI provider.

The foundation is two test suites: browser tests using Playwright and regression tests coordinated across several AI agents by a control program, with documentation in Xray. The tools check core workflows and evaluate test accounts and test data in real time, helping catch problems before they reach production.

Where I can help

I introduce AI agents into project and testing teams: defining tasks, setting boundaries, agreeing acceptance criteria and bringing the team along. The principle is speed within clear safeguards. Let’s discuss working together.

This website as an example

This website was built in the same way. An agent produced the first version in one night, with backups, checksums and independent checks; the previous WordPress site is now archived. Since then, I have developed it further over several rounds with Codex and Claude Code. I set the goals, content and direction, choose between proposals, review, correct and approve. The tools provide drafts and implementation. The portraits are AI-generated from my own photographs and labelled accordingly.

I write things down. Clear responsibilities, recorded decisions and direct communication matter to me. I address disagreements fairly and on their merits. I enjoy working with people who think for themselves, and helping them develop.

Structure and room to act

Clear responsibilities give a team room to make its own decisions. Recording decisions keeps knowledge accessible without prescribing every action. Automated checks free up time for situations that need human judgement.

My aim: what I build should keep working without me.