Dream Machines
Dream Machines provides robotic AI systems that learn manufacturing tasks directly from operator demonstrations, enabling European SMEs to automate variable, repetitive workbench work without programming expertise. The robots train overnight, improve continuously from their own operational experience, and can be reconfigured the same day when products change. The system targets semi-structured tasks like PCB testing, battery slotting, and machine tending with variable parts.
- Robotics
- Manufacturing / Industry 4.0
- Industrial Automation
- Artificial Intelligence
- Hardware
Funding
Founders
Founder details are not available yet.
Product
Problem
European manufacturing SMEs rely on skilled operators for variable, repetitive workbench tasks that are essential but too inconsistent for traditional automation. With labor scarce, costs rising, and experienced operators retiring, these companies struggle to scale production without proportional increases in headcount.
Solution
Dream Machines provides robotic AI packaged as a full solution that integrates into existing workstations and learns directly from operators through demonstration. Instead of programming every movement, the person who knows the task best demonstrates it, and the robot learns overnight, clocks in the next morning, and keeps improving from its own experience with every shift it works. When products or processes change, the automation is updated the same day through new demonstrations. The system focuses on semi-structured tasks such as PCB testing, battery slotting, machine tending with variable parts, picking and sorting irregular objects, and quality control—critical work that rarely defines strategic advantage but is exactly where intelligent robotics creates leverage.
Target Audience
Primary customers are European manufacturing SMEs with a few hundred employees that are global leaders in niche markets and need to automate variable, repetitive workbench tasks without hiring robotics specialists.
Features
- Demonstration-based learning that captures operator expertise without requiring programming skills
- Overnight training cycle that prepares the robot for the next workday automatically
- Continuous improvement from the robot's own operational experience on each shift
- Same-day reconfiguration through new demonstrations when products or processes change
- Full-solution integration into existing workstations without external integrators
- Targeting of semi-structured tasks like PCB testing, battery slotting, machine tending, pick-and-sort, and quality control
- Built on fine-tuned open-source models, with published empirical results from real manufacturing deployments