Humanoid robots are leaving the research lab. The hardware has become capable and affordable, and the question is no longer whether a humanoid can move but whether it can do useful work, safely, alongside people.

At ICT Strypes Portugal, we build the software layer that decides this: the skills a robot learns, the data it learns from, and the evidence that it performs reliably. Our Physical AI team works hands-on with a full-size humanoid robot in our Portuguese lab, together with ICT Group's Center of Innovation in the Netherlands. What we develop in Portugal runs on robots across the Group.

From Programming to Teaching

Programming a robot used to mean writing every motion by hand. Today, robots learn from demonstration. An engineer performs a task, the robot records it, and an AI model learns to repeat it, even in situations it has never seen. Our engineers combine this new way of working with deep experience in embedded software.

Our expertise includes:

  • Teleoperation and demonstration capture using VR and motion-capture gloves
  • Dataset curation, labeling, and quality assurance
  • Fine-tuning of Vision-Language-Action (VLA) foundation models
  • Simulation and digital twins for synthetic training data
  • Integration on edge AI platforms such as NVIDIA Jetson

Safe, Measurable, and Ready for the Workplace

A humanoid robot is machinery, and in Europe it will be deployed under the EU Machinery Regulation from January 2027. 
We build safety in from the first experiment, and we measure performance in numbers rather than in demos. Every task we take on produces a lasting result: a dataset, a trained model, an evaluation report, and a safety file.

Key capabilities:

  • Use-case selection and robot readiness assessments
  • Evaluation in simulation and on the real robot: success rates and failure analysis
  • Safety cells, risk assessment, and compliance documentation
  • Secure, isolated robot networks
  • An open, vendor-neutral data format, so your investment is not tied to one robot or one model

Shaping Physical AI in Europe

We don't aim to be the biggest; we aim to be the team that knows how to do this well. 
Do you have a task a humanoid robot could take on in your workplace? We are selecting partners and use cases for our first pilots.
Let's explore it together.

More information?

Please contact our Managing Director, Geert Eijkhout.

Send an email Make a connection
Geert Eijkhout