# Interview guide: Alpay Celik

You are helping someone prepare a job interview with Alpay Celik. Below is what his portfolio says about him (the same text as https://alpaycelik.dev/llms.txt), followed by the job ad if there is one.

1. Write 8 to 10 interview questions. With a job ad, aim them at its requirements, the most important first. Without one, aim them at a senior full-stack role that builds with AI agents.
2. For each question give: the question itself; what in the portfolio it builds on, with a link to that place; and what a strong answer would cover, in two or three short points.
3. Use only the portfolio text below as evidence and only URLs that appear in it. Never invent a project, a skill or a link.
4. Add two questions on requirements of the ad that the portfolio does not show, and say plainly that they are gaps to check in the interview. Without an ad, skip this step.
5. Prefer questions that make him explain a decision and its alternative over questions that can be answered with a yes.
6. Write the whole guide in the language of the job ad, or without one, in the language of the person asking.
7. If the job ad is only a link, open it and work from its text. If you cannot open it, say so and continue without the ad.
8. End with the contact form, https://alpaycelik.dev/#contact.

==================== PORTFOLIO ====================
# Alpay Celik

> AI fullstack engineer based in Germany. How I work with agents, selected projects, and the practice behind them.

AI fullstack engineer based in North Rhine-Westphalia, Germany. I work the whole distance, from the data model through the API to the last pixel - and most of it now gets built with agents: spec first, gates in between, evidence at the end. Language models are both the tool and a part I build with. What comes out has to be checkable, not just plausible.

The site is bilingual: German is the default at https://alpaycelik.dev, English is at https://alpaycelik.dev/en, and every page exists in both. The links below point at the English version.

What I work with, by area:

- Agentic software engineering: Claude Code, Claude Skills, MCP, Multi-Agent Orchestration, Context Engineering, Python
- Frontend: TypeScript, React, Next.js, Angular, React Native, Redux, Signals, TanStack Query, Tailwind, SASS
- Backend & data: Node.js, Express, .NET, C#, Python, MongoDB, PostgreSQL, SQL, Prisma, Swagger
- Infrastructure & process: Git, GitHub Actions, GitLab CI/CD, Azure DevOps, Docker, Playwright, TestCafe, Jira, Confluence
- Design fluency: Figma, Adobe XD, Design Tokens, A11y

## Projects

- [MCP Migration Check](https://alpaycelik.dev/en/projects/mcp-migration-check): The MCP 2026-07-28 revision breaks nearly every existing server. One rule engine as a web demo, npx CLI, GitHub Action and agent skill. Stack: TypeScript, Python, Rust, Go, Next.js, Cloudflare Workers, node:test, GitHub Actions. Live: https://mcp-migration-check.alpaycelik.workers.dev. Source: https://github.com/AlpayC/mcp-migration-check.
- [Abfall MCP](https://alpaycelik.dev/en/projects/abfall-mcp): Roughly 400 waste authorities, each with its own portal, no nationwide API. An MCP server that finds the right one among 995 from an address. Stack: Python, MCP, Streamable HTTP, Nominatim, pytest, Docker, GitHub Actions. Live: https://abfall-mcp.alpaycelik.dev. Source: https://github.com/AlpayC/abfall-mcp-server.
- [Claude Skill Set](https://alpaycelik.dev/en/projects/claude-skill-set): 34 skills for Claude Code in enterprise codebases. Autonomy comes from moving the human decision out of the middle and to the edges. Stack: Claude Code, Agent Skills, Hooks, Node.js, PowerShell, Markdown. Source: https://github.com/AlpayC/claude-skill-set.

## On the site

- [How I work with agents](https://alpaycelik.dev/en#agents): A substantial share of the code I deal with every day is now written by agents, often several in parallel, each with its own task and its own context. That doesn't take the work away, it moves it from writing to checking - where more code gets produced, reading is what becomes expensive. So a run hands over more than a diff: gates green, screenshots of the states, and the decisions it took on its own along the way. Five things I've learned doing it. Principles: Context beats prompting; Verifying is the job; Small steps; Scope the work; Learn to push back.
- [How an idea becomes a deploy](https://alpaycelik.dev/en#process): Four steps I keep the same, whether it's a feature, a prototype, or a side project in the evening. An agent works inside each of them now, which changes nothing about the order: surprises at the end cost more than questions at the start.

## Contact

- [Email](mailto:hello@alpaycelik.dev): A technical question, a disagreement with something I wrote, or just a hello. Write to me. I read everything and reply as soon as the day allows.
- [GitHub](https://github.com/AlpayC): The source of the projects above, and more repositories
- [LinkedIn](https://www.linkedin.com/in/alpay-celik/)

## Optional

- [MCP server](https://alpaycelik.dev/mcp): The same content as tools for agents, over Streamable HTTP (MCP 2026-07-28). Add it with `claude mcp add --transport http alpay https://alpaycelik.dev/mcp`
- [German version](https://alpaycelik.dev): The same site in German, the default language
- [System status](https://status.alpaycelik.dev): Live uptime monitoring
- [Sitemap](https://alpaycelik.dev/sitemap.xml): Every URL in both languages

<<< Paste the job ad below this line >>>