About Promethion Industries

Promethion Industries is a European defence technology company building autonomous supremacy systems for aerial and multi-domain operations. We design tightly integrated autonomy stacks in Europe to enable decisive autonomy for NATO-aligned forces, focusing on defensive and counter-system capabilities including anti-UAS and resilient autonomy under contested conditions. Alongside this, our team explores open code libraries and hardware platforms for robotics vision and perception A.I. powered by fully synthetic data, challenging corner paradigms in A.I. development to make autonomous robotics more accessible.

Your mission with us

• Research and develop event-based perception for high-speed airborne target detection and tracking.

• Develop asynchronous feature detection, tracking and target-motion estimation using event-camera data.

• Explore event-native neural networks, recurrent models, graph-based approaches and spiking neural networks.

• Build multimodal fusion approaches across event cameras, RGB and thermal sensing.

• Evaluate when event-based sensing genuinely outperforms conventional frame-based pipelines.

• Optimize event processing for ultra-low latency and edge deployment.

• Work with Prophesee and similar event-camera hardware.

• Build real datasets from flight testing and quantify performance under high dynamics, low texture, HDR and rapid closure.

• Transfer validated components into the production perception stack.

The current resource base spans asynchronous target tracking, combined frame/event detection and multiple event-based tracking approaches.

Your contribution to the team (required)

• Strong background in event-based vision, neuromorphic sensing or asynchronous computer vision.

• Strong computer-vision and machine-learning fundamentals.

• Experience implementing research algorithms beyond offline notebooks.

• Strong Python and C++ skills.

• Experience working with real sensors and raw sensor data.

• Strong experimental methodology and ability to quantify latency, robustness and accuracy.

Even if you don’t tick every box, we encourage you to apply - we’d love to hear from you.