Available for new projects · Stuttgart

AI that actually arrives in your everyday work.

We analyse where AI actually creates leverage in your business or home — and then build it pragmatically. No slide decks, just working solutions in weeks instead of months.

Stuttgart & Region Reply within 24h Clear project scope
AI Assistant
How do I automate our invoice review?
I can extract, verify and post incoming invoices in your system.
Living Room
Lights
Heating21.5°C
Shutter
BTC/USD · 1H
67 420+2.4%
Live from Stuttgart
Experience 20+ years Hardware & Software
Focus KI · Smart Home · E-Commerce · Trading
Models GPT · Claude · lokale LLMs (Ollama)
Languages DE · EN · RU · ZH
Request initial call Reply within 24h · Stuttgart

AI is changing everything. Now.

Those who engage with Artificial Intelligence today will be indispensable tomorrow. Those who don't will be replaceable.

We help you stand on the right side.

Smart home interior with holographic AI visualisation
Hardware meets software meets AI

Where others show PowerPoint slides,
we deliver working systems.

Services

Four areas, one principle:
less talk, more delivery.

We only take on projects when we're sure AI will give you real leverage. If not, we'll tell you straight.

Smart home components and wiring

Home Automation

Smart home consulting and setup — from first sketch to running system. With devices that stay under your control.

  • Planning, device selection, wiring
  • Home Assistant, Matter, Zigbee, KNX
  • Automations & scenes (light, heating, audio)
  • Alarm & security systems
  • Remote maintenance & 24/7 support options
Typical scope
2–8 weeks
Collaboration
Project-based or ongoing
AI model integration and automation

AI Services

Intelligent automation and AI agents for your business — from chat assistants to fully automated processes.

  • AI chatbots & assistants (internal + customer-facing)
  • Document processing & data extraction
  • Business process automation (RPA + LLM)
  • Custom AI solutions with your own data
  • AI strategy & roadmap consulting
Typical scope
4–12 weeks
Collaboration
Concept, build or operation
Trading bot visualisation with market data

Algorithmic Trading

Development and operation of AI-powered trading bots for stocks and crypto — as a purely technical service.

  • Trading bot development (Python, Rust)
  • Backtesting & strategy optimisation
  • Broker integration & order automation
  • Performance monitoring & risk limits
  • Own infrastructure (no cloud lock-in)
Typical scope
6–16 weeks
Collaboration
Technical implementation
SHOP ERP LAGER BUCH. Webshop and ERP integration

Webshop & Integration

Webshop development with real bidirectional integration to your accounting and inventory system — no more double entry.

  • Webshop build (Shopify, WooCommerce, custom)
  • Bidirectional accounting integration (DATEV, lexoffice, sevDesk)
  • Inventory sync (stock, incoming, shipping)
  • Automatic invoicing & receipt generation
  • Marketplace connectors (Amazon, eBay, Otto)
Typical scope
4–12 weeks
Collaboration
Project-based or ongoing
Note: In algorithmic trading we offer purely technical services — development, backtesting, operation. This is not investment advice, asset management or portfolio management. Strategy and risk remain entirely with you.
Approach

Four steps — and you'll know where you stand.

We work in short, visible iterations. After each step you'll have a concrete result in hand and can decide whether to continue.

Step 01

Discovery

We listen. Understand what you do, where it hurts, which AI levers are realistic and which aren't.

Result · 1 workshop, 1 memo
Step 02

Concept

Mapping use cases against effort & value. You get a prioritised roadmap with clear next steps.

Result · Roadmap + offer
Step 03

Build

We build the MVP in weeks, not months. You see progress early — and can change direction.

Result · Working system
Step 04

Operate

We stay on it — monitoring, updates, extensions — as long as you want. Or you take over yourself.

Result · Stable operation
Use Cases

What it looks like in practice.

Three typical scenarios from our work. Concrete enough that you'll recognise yourself — abstract enough not to breach any client confidentiality.

Use case: contract analysis for law firm
Law firm · 12 employees

Classify contract clauses in seconds instead of hours.

An LLM agent reads uploaded contracts, flags risk clauses and suggests alternative wording from the firm's own clause library.

Effect Review time −70%
Use case: architecture chatbot
Architecture firm · 8 employees

AI chatbot that pre-qualifies incoming project inquiries.

An assistant asks follow-up questions on the plot, budget and timeline, and produces a pre-sorted briefing for the architect — before the first conversation takes place.

Effect +40% lead quality
Use case: crypto backtest platform
Private investor · Crypto portfolio

Backtest platform for your own trading strategies.

Strategies are tested against 5 years of historical data with risk metrics and Sharpe-ratio analysis. Delivered as a maintainable Python repository.

Effect 200+ strategies tested
About

20 years of hardware & software.
No sales team.

Behind dein-ki.haus is Evgeny Popov — with over 20 years of hands-on hardware and software. From smart-home wiring to webshop backends, from ERP interfaces to AI agents. Based in Stuttgart, with clients across DACH.

We're deliberately small. That means: you work directly with the person who builds — not an account manager who vanishes after the sale. You don't get a 100-page strategy, you get a working system. And if a project is too big for us, we'll say so.

Our promise: honest assessment, deliverable results, transparent collaboration. If AI isn't the right path for you, we'll say so — at the very first call.

Location Stuttgart
Languages DE · EN · RU · ZH
Stack Shopify · DATEV · Home Assistant · LLMs
Models GPT · Claude · Llama · Mistral
FAQ

What clients ask us before the first call.

How long does a typical AI project take?
Depends on scope. A chatbot with its own knowledge base: 4–6 weeks. A document pipeline integrated into your existing system: 6–10 weeks. Smart-home setup: 2–4 weeks. After the first call you get a binding timeline — no "it depends" answers.
What happens to our data?
Sensitive data stays with you. We can run models locally (Ollama, vLLM) or use GDPR-compliant EU endpoints (Azure OpenAI EU, Mistral, Aleph Alpha). For standard workloads without personal data we also use OpenAI/Anthropic directly. You decide what runs where — we advise honestly on what makes sense.
Do we need our own servers or cloud infrastructure?
No. Most projects run on existing infrastructure, in the cloud (Hetzner DE, AWS Frankfurt) or hybrid — your choice. If you have strict on-premise requirements, we'll set that up. We avoid vendor lock-in by default.
Which AI models do you use?
We're model-agnostic. For most tasks Claude 4.5, GPT-5 and Mistral Large are first choice. For local / privacy-critical setups: Llama 3, Mistral, Qwen via Ollama. For very specific tasks we also train small custom models. Which models fit your case is part of discovery.
How is the scope clarified?
In the initial call we clarify the goal, current setup, data, systems and risks. Afterwards you receive a written assessment with a sensible approach, realistic scope and the next decision points.
Can we outsource just a small part?
Yes. We also take small, clearly scoped engagements — a strategy session, a code review of your existing bot, or a prompt-engineering workshop for your team. We clarify the right starting point together.
What if AI isn't a fit for our problem?
We tell you in the first call. Honestly: in roughly a third of inquiries, classical automation (scripts, Zapier, Make) is the better answer than an AI model. In that case, we recommend exactly that route.
What sets you apart from a big consultancy?
Three things: 1) You talk directly to the person who builds. 2) We deliver working code, not PowerPoint slides. 3) We say "no" when AI isn't the right lever. What often gets stuck in months of coordination inside large consulting structures, we approach leanly and directly. In return we're not the right fit for 200-person enterprise transformations.
Initial Call

Let's have a quick chat — direct and non-binding.

30 minutes. You tell us what you're up to. We say whether AI fits, what scope is realistic and what the next sensible step looks like. Nothing more.

Reply within 24h No sales pitch Written assessment afterwards