Home / Custom software development

Custom software development

This is where we are truly at home. We have worked on a wide range of projects, from multi-year enterprise development to experimental R&D.

Long-term projects

Complex development in large teams

Development over many years with teams of 15-20 people, typically for German and US clients. We know what it takes to build a large system that stays sustainable: architecture, code quality, continuous delivery, documentation and working closely with the client's team.

R&D

Machine vision and AI training data

Smaller, focused R&D work. Before the LLM boom, with 40-50 annotators, we collected and checked training data for AI, typically in machine vision, and we also verified the training.

What we do

With experience from about 30 different software projects.

  • Design and development of web and cloud applications
  • Further development and takeover of existing systems
  • Data collection, annotation and AI training data
  • Software that talks to PLCs, sensors and machines
  • Operation, monitoring, CI/CD
  • Documentation in English, German or Hungarian

Technology: a tool, not the goal

We have worked in C#, .NET, Angular, React, Python and C, and we are at home in Microsoft Azure. But today knowing a particular language is not what makes the difference: we use AI tools intensively in development, so we always pick the technology to fit the task. What matters is understanding the problem, good architecture and a working, maintainable system.

C# / .NETAzureAngularReactPythonCAI-assisted development

How we work

We never start without a plan. Our work always leaves a written trail and stays traceable and transparent.

  1. UnderstandingGoals, users, existing systems.
  2. PlanWritten plan with estimate and schedule.
  3. DevelopmentShort cycles, regular demos.
  4. HandoverDocumentation, deployment, training.
  5. OperationMonitoring, fault alerts, further development.
  6. Our own productsWe build our own EMP, AH and MMC systems the same way.
Featured project, 2020-2023

3D scanning and automatic damage detection

In an international product development we built high-resolution, photogrammetry-based 3D models of vehicles with texture mapping, and made the data usable in the real world through augmented reality (AR).

The goal was to detect damage on the car body automatically and classify it by severity. For this we trained our own AI model on many terabytes of training data, before the LLM boom.

Photogrammetry3D meshTexture mappingARMachine visionOwn AI model

3D scanning and damage detection

Looking for a development team?

Tell us briefly what you are working on, and we'll get back to you soon.

Get in touch