Websites & small web projects
Clean, fast websites for projects, independent businesses, associations, or internal initiatives — focused on clarity, structure, and easy maintenance.
$ services
AS
Andre Semek · Berlin
Websites, automations, and pragmatic AI workflows — built clearly and carefully.
I help small projects, independent teams, and organizations turn technical ideas into useful work. My background is cloud, DevOps, and infrastructure — which means I care about what happens after launch: clarity, maintainability, fewer moving parts.
Small projects are welcome. A short email is enough — no polished brief needed.
Scroll
$ what-i-help-with
Clear offers for well-scoped work
Automations
Practical workflow automation with tools, APIs, scripts, or orchestration platforms. Less repetitive work, fewer avoidable handoffs — without complexity for its own sake.
Pragmatic AI workflows
Lightweight AI-supported workflows for research, documentation, preparation, and internal helpers — designed with review, guardrails, and handover in mind.
Technical sparring
A calm second brain for cloud, DevOps, infrastructure, and delivery questions — when you need to make the next step clearer, safer, or more maintainable.
about me
$whoami
Berliner. Football fan. Music lover. Father of two.
15+ years in infrastructure, automation & operations.
Currently: Cloud Engineer / DevOps / Solution Designer @ Vattenfall.
15+ years in infrastructure, automation & operations.
Currently: Cloud Engineer / DevOps / Solution Designer @ Vattenfall.
$cat ~/philosophy.txt
I like working in roles where technical work concretely benefits others.
Automation is also a hobby — my homelab runs things I actually use daily, not just test.
Since working with AI agents, my daily workflow has fundamentally changed.
Automation is also a hobby — my homelab runs things I actually use daily, not just test.
Since working with AI agents, my daily workflow has fundamentally changed.
$focus
→ Solution design for real operational problems
→ Azure, Azure DevOps, CI/CD Pipelines
→ Bicep & Terraform for maintainable IaC
→ Linux-first operations & CLI workflows
→ AI-supported automation (n8n, AI Agents)
→ Azure, Azure DevOps, CI/CD Pipelines
→ Bicep & Terraform for maintainable IaC
→ Linux-first operations & CLI workflows
→ AI-supported automation (n8n, AI Agents)
$_
Skills & Tools
Cloud & Infrastructure
Azure
Azure DevOps
Bicep
Terraform
Docker
Kubernetes
Linux
Proxmox
Vault
Traefik
TailscaleAutomation & CI/CD
GitHub Actions
Jenkins
Bash
PythonCI/CD
AI Agents
n8n
Git
AnsibleMonitoring & Observability
Prometheus
Grafana
PostgreSQL
ELKAlso comfortable around HAProxy, OPNsense, TrueNAS, Zabbix, and a grown homelab where new setups are not just tested but actually used daily.
AI, Automation & Development
Since early 2025 I use AI agents daily: for planning, reviewing, debugging, documentation, and workflows in my homelab and at work. I have spent over a year figuring out what actually works — and what doesn't.
AI Assistants
Claude
ChatGPT
Gemini
GitHub CopilotDSDeepSeekKiKimiMMMiniMaxGLMGLMOCOpenCode
Cursor
WindsurfDevelopment
VS Code
Warp
ZedCLClineKLKilo CodeTRTraeRCRooCodeWeb Design
v0.devCAClaude Artifacts
Lovable
FigmaAutomation
n8nOCOpenClawAIAI AgentsPEPrompt EngineeringTools & Models
Editorial
AI in everyday work
Working with AI — pragmatic, safe, team-ready
For me, AI is genuinely helpful when a ticket becomes a traceable process: gather context, research, plan, implement, verify, document, hand over. That is where real value emerges in daily work.
I see AI not as a replacement for engineering, but as an amplifier for clean working practices. Those who work in a structured way gain speed and clarity. Those who work messily mainly scale confusion.
Context before promptVerification before trustTeam-readiness before solo hacks
Which AI tools are actually useful
The most important question for me is not which tool is being marketed loudest right now, but which setup fits the task, the team, and the security framework.
Why AI without guardrails gets expensive fast
Secrets, screenshots, internal systems, sensitive docs, and unclear data flows are not edge cases for me. Those who address this later build uncertainty directly into the workflow.
What actually scales with agents in a team
Individual experiments can quickly seem productive. But it only becomes sustainable when prompts, instructions, roles, reviews, and verification become understandable to others.
How AI output becomes reliable
Speed helps little if the result is unclear or wrong. I treat AI output like code in review: check it, question it, verify it — before it moves on.
My starting point is almost always the same question: How can a team use AI in a way that creates less friction, better decisions, and more traceable technical work?
From practice
This is less about buzzwords and more about working style, current topics, and the kind of improvements I typically work on in practice.
15+ yearsInfrastructure, operations, and automation.
TodayCloud Engineer / DevOps / Solution Designer at Vattenfall.
Working styleCalm, structured, pragmatic, and reliable.
Making pipelines calmer
I enjoy working on setups where releases become more traceable, quality is visible earlier, and teams have less friction in their daily work.
Keeping IaC maintainable
Bicep, Terraform, and clean standards help me not just provision infrastructure, but keep it maintainable over time.
Integrating AI meaningfully
I deploy agents and workflows where they bring real relief: analysis, documentation, preparation, and recurring tasks.
$ selected-work
Selected work — shown as compact case snapshots
Intentionally concise. The goal is to show what kind of work this is and why it matters — not to inflate it.
SlimUtils
A browser-based collection of practical everyday utilities, packaged in a lightweight product experience that prioritizes speed, clarity, and low friction.
$ what it shows Information architecture for a growing tool set, clean utility UX, and a bias toward practical value over feature theater.
Pridovo
A mobile-first prediction league product with a clearer matchday flow, simpler rankings, and a more modern feel than older desktop-heavy tip-game experiences.
$ what it shows Product thinking, mobile-first interaction design, and a cleaner way to structure recurring sports data and social competition.
Kila Combo
An information-rich site for a Kreuzberg parent-initiative daycare, structured to help families and applicants quickly understand the offer, team, and day-to-day setup.
$ what it shows Content structure, approachable navigation, and practical communication for a community-oriented organization.
Marie Jacob
A visual shop and editorial-style web presence for sustainable interior and lifestyle products, balancing browsing, atmosphere, and straightforward navigation.
$ what it shows Visual hierarchy, calm e-commerce presentation, and a softer content-to-shop transition.
Eurocoeff
A sports-data project that makes football coefficient rankings, comparison views, and historical context easier to browse and understand.
$ what it shows Data presentation, comparison UX, and a preference for making specialist information easier to scan.
Open to project inquiries, freelance work, and good tech conversations.
Contact
If your context touches websites, automations, cloud, DevOps, or AI-supported workflows, I look forward to the conversation. Small projects are welcome — a short note is enough.
andre@semek.org