BERLITEC

Software development · AI consulting · Berlin

Software, AI and agents – from the idea to production.

BerliTec GmbH has been building custom software since 2001 and advises companies on artificial intelligence, large language models, locally hosted models and autonomous agents.

Discuss your projectSee services

Services

What we do for you

We work at the intersection of product, data and model – from the first analysis to the running system.

  1. 01

    Consulting & technology advisory

    We assess where software and AI create real value – and where they do not.

    • Feasibility and opportunity assessment
    • Technology choice: cloud, hybrid or local
    • Architecture, data flow and interfaces
    • Cost, operations and scaling models
  2. 02

    Software development

    We build web applications, backends and interfaces that hold up in daily use.

    • Web frontends and backends
    • APIs and databases
    • Mobile and desktop applications
    • Integration into existing systems
  3. 03

    AI and large language models

    We integrate language models into products and processes – controlled, measurable and affordable.

    • LLM integration into existing applications
    • Retrieval-augmented generation (RAG) over your data
    • Prompt, testing and evaluation strategy
    • Cost control and quality assurance
  4. 04

    Local models & data protection

    When data must not leave the building, we run models where the data lives.

    • Operation with Ollama, llama.cpp or vLLM
    • On-premises and private cloud
    • No transfer of your data to third parties
    • GDPR-compliant processing
  5. 05

    Agents & automation

    We build agents that do more than answer: research, extract, verify, trigger.

    • Agents with tools (tool use)
    • Multi-step workflows and approvals
    • Automation of recurring processes
    • Guardrails, logging and control

Focus

Artificial intelligence that lands in the business.

Large language models are not an end in themselves. We bring them to where they take work off your plate: into your processes, with your data, under your responsibility.

  • Local and open models on your own hardware – with no data leaving the building
  • RAG over your documents, databases and file stores
  • Agents with tools for research, extraction and process steps
  • Evaluation, guardrails and monitoring in live operation
Local operation, two commands
$ ollama serve
$ ollama run qwen2.5:7b-instruct
# runs entirely on your hardware – no data leaves the building

How we work

Four steps, no detours

  1. 01

    Understand

    We clarify the goal, the constraints and the data – and what the result has to deliver.

  2. 02

    Prototype

    Within a short time there is a working prototype you can use to test decisions.

  3. 03

    Build

    We take the solution to production, with tests, error handling and clear documentation.

  4. 04

    Operate

    We support the launch, monitor quality and keep developing the system.

Technologies

What we work with

We choose tools by the task, not by fashion – and rely on traceable, maintainable systems.

Languages & frameworks

  • Python
  • TypeScript
  • JavaScript
  • Node.js
  • FastAPI
  • React

Data

  • PostgreSQL
  • SQLite
  • Vector databases
  • Redis

AI

  • Ollama
  • llama.cpp
  • vLLM
  • Hugging Face
  • OpenAI-compatible APIs
  • RAG

Operations

  • Docker
  • Linux
  • Nginx
  • CI/CD
  • Monitoring

Platforms

  • Web
  • iOS
  • Android
  • Desktop

Contact

Let's talk about your project.

Send us a short message. We will get back to you with a first assessment.

Get in touch