Machine Intelligence Zrt.

Our journey began with a fundamental insight: the same AI technologies that can detect nanometer-scale defects in semiconductor manufacturing can also identify potentially hazardous asteroids in astronomical images, and the same systems that optimize industrial facility management can enhance spacecraft navigation. This understanding has shaped our dual-focus approach, in which capabilities developed in one domain continuously strengthen our offerings in the other. The precision required by aerospace manufacturers teaches us about reliability and traceability, while the data scarcity challenges of space applications drive our innovation in synthetic data generation and edge computing.

In the space domain, our work extends across multiple critical areas. We are actively executing projects with the European Space Agency, including our flagship NEODetect system developed in collaboration with ESA's Planetary Defense team. NEODetect applies deep learning to analyze observatory astrophotography in near real-time, identifying potentially hazardous near-Earth objects with the speed and accuracy required for effective planetary defense. This system exemplifies our approach to space applications: taking proven AI architectures, adapting them to the unique constraints of space data, and delivering operational systems that address real mission needs.

Equally significant is our DLVS3 (Deep Learning Visual Space Simulation System), a proprietary platform that addresses one of the most persistent challenges in spacecraft autonomy: the scarcity of annotated training data for critical operations. Spacecraft approach maneuvers, docking procedures, and coordinated swarm movements have historically lacked the extensive image databases needed to train robust AI systems. DLVS3 solves this by using physics-based simulation to generate millions of realistic, annotated samples that incorporate accurate celestial mechanics, spacecraft parameters, and sensor characteristics. The system maintains connections with NASA JPL's SPICE framework, ensuring astronomical precision in the synthetic data it produces. This capability is proving essential as the space industry moves toward industrial-scale operations requiring automated systems with proven reliability.

Our work in spacecraft Guidance, Navigation, and Control represents another frontier where simulation meets operational AI. We are developing visual-based deep learning solutions for satellite positioning within GNC systems, supported by DLVS3's ability to generate the diverse training datasets these applications demand. The integration of computer vision with traditional GNC architectures enables new approaches to relative navigation, autonomous rendezvous, and formation flying, all critical capabilities for the emerging satellite constellations and on-orbit servicing missions.

At the heart of our capabilities lies the MInD (Machine Intelligence Designer) platform, a comprehensive AI development suite that has been deployed across six countries in demanding industrial environments. MInD embodies our philosophy of practical, hardware-agnostic artificial intelligence that adapts to existing infrastructure rather than requiring wholesale system replacement. The platform supports the complete lifecycle of AI development, from multi-user collaborative annotation to model training to edge deployment, with particular strength in visual and time-series data applications. Our experience spans advanced measurement techniques, including ellipsometry, white-light interferometry, and Raman spectroscopy, demonstrating our ability to work with the sophisticated instrumentation that characterizes both precision manufacturing and space science.

Core Capabilities:

• Deep learning for visual and time series analysis

• Spacecraft GNC and autonomous navigation

• Planetary defense and NEO detection

• Synthetic training data generation (DLVS3)

• Edge computing and space-rated AI systems

• Industrial quality control and precision manufacturing

• Computer vision and advanced measurement integration

Seeking Partnerships In:

• Consortium collaborations for ESA and EU space programs

• Autonomous spacecraft operations and on-orbit servicing

• Space manufacturing and quality control systems

• Edge computing for satellite constellations

• Planetary defense and space domain awareness

• Academic research collaborations

• Technology development grants and international initiatives