Job Description:
We are seeking a Computer Vision Engineer (AI Perception) to design and implement perception systems that enable ROZOR robots to understand complex indoor environments.
You will work on algorithms for detection, segmentation, 3D mapping, tracking, and multi-sensor fusion, ensuring that ROZOR robots can interpret dynamic, cluttered spaces with high accuracy and robustness.
Your work directly impacts how ROZOR robots perceive people, obstacles, and mission environments—improving their safety, autonomy, and operational intelligence.
Responsibilities:
- Develop computer vision and deep learning models for:
- Object detection and 2D/3D tracking
- Semantic and instance segmentation
- Depth estimation, scene reconstruction, and mapping
- Implement multi-modal fusion between camera, LiDAR, depth sensors, and IMUs
- Build real-time perception pipelines optimized for embedded GPUs (Jetson)
- Analyze sensor data, identify failure cases, and improve robustness
- Integrate perception modules with ROS/ROS2 and the autonomy stack
- Conduct field testing to validate perception performance on ROZOR robots
- Collaborate with autonomy, embedded, and hardware teams to optimize camera and sensor configurations
- Maintain datasets, labeling workflows, and model training infrastructure
Preferred Qualifications:
- Bachelor’s/Master’s degree in Computer Vision, AI, Robotics, Computer Science, or related field
- 2+ years experience building computer vision or perception systems
- Strong proficiency with Python and deep learning frameworks (PyTorch or TensorFlow)
- Experience with image processing, feature extraction, and multi-view geometry
- Hands-on experience with detection/segmentation models (e.g., YOLO, Mask R-CNN, DeepLab, DETR)
- Experience working with LiDAR or depth sensors
- Familiarity with ROS/ROS2, OpenCV, and perception debugging tools
- Ability to optimize perception algorithms for real-time inference