Skip to content
Portrait of Jan Niklas Kreimeier

Jan Niklas
Kreimeier

DevOps & Cloud Engineer · Full-Stack Web Developer

about

I'm a full-stack developer passionate about building modern, maintainable, and scalable web applications. My expertise spans frontend development, backend architecture, and cloud infrastructure, with a strong focus on creating clean, reusable, and secure software.

skills

My specialization lies in DevOps, where I design and build scalable delivery platforms that enable teams to develop, test, deploy, and operate software efficiently. I focus on Infrastructure as Code (IaC), CI/CD automation, release management, workload monitoring, and production operations, creating reliable delivery pipelines that improve consistency, scalability, and developer productivity.

~/jk @ ops
$
ai

I work fluently with AI across the full spectrum of automation, augmentation, and autonomous agents. I leverage AI to automate repetitive tasks, collaborate on complex problem-solving and software development, and orchestrate AI agents that operate independently within well-defined objectives. My approach emphasizes effective, ethical, and secure AI integration to build practical, production-ready solutions.

  • Claude Codedaily driver

    Claude with the everything-claude-code plugin runs in my terminal: multi-file refactors, test-first changes, repo-wide reviews. I read every diff before it lands.

  • GitHub Copilotin-editor

    Inline completions and chat where the context is already loaded. The boring 80% so I can focus on the part that actually needs thinking.

  • ChatGPTthinking partner

    Architecture sketches, rubber-ducking trade-offs, drafting docs and migration plans. Less code, more clarity.

  • Lovableprototyping

    I use it for drafting UI ideas and quick proof-of-concepts. Good for exploring layouts fast, then rebuilding the keepers properly in my own stack.

delegation

I decide what work to hand off to AI and what needs my own hands on the keyboard. Repetitive scaffolding and refactors go to the agent; architecture, trade-offs, and anything touching production data stay with me.

description

Good prompts are a skill. I write clear context, constraints, and expected output formats. The better I describe the problem, the less time I spend fixing the result.

discernment

I treat every AI output as a first draft. I read every diff, check for hallucinations, and run the same CI pipeline on agent-authored code as I do on mine. Quality control never delegates.

diligence

Scoped contexts for anything touching customer data. No production credentials in chat windows, ever. Responsible AI collaboration means the same security standards apply whether the author is human or machine.

minigame
assistant idle

The AI will stream patch suggestions. Approve the safe ones, reject the unsafe ones — hallucinated APIs, leaked secrets, missing auth, silenced tests. Press start.

Mocked assistant — every diff is hand-written. The point is the review reflex: read it, then decide.