Computer Vision Engineer

LUXUAV is an innovative technology company headquartered in Luxembourg. We specialize in the design, development, and manufacturing of advanced Unmanned Aerial Vehicles (UAVs). Our modular systems are tested in real-world conditions and specifically tailored for both commercial and security applications.

Role Overview

Develop and optimize perception algorithms for UAV platforms. Focuses on real-time detection, classification, tracking, vision-based pose estimation, EO/IR sensor fusion, and edge deployment under strict latency, power, and memory constraints in safety-critical, defense-grade environments.

Key responsibilities

Computer Vision & Perception:

• Design and implement real-time object detection, classification, multi-target tracking, and change detection for autonomous UAV operations, including high-altitude ground target detection, camouflage, and foliage occlusion.
• Optimize deep learning models (YOLO, transformers, CNNs, or proprietary architectures) for edge deployment on embedded platforms (NVIDIA Jetson, Qualcomm RB5) under strict latency, power, and memory budgets.
• Develop vision-based pose estimation, optical flow, and feature tracking pipelines for mission-critical applications.
• Implement camera calibration routines and multi-camera rig setups for stereo and wide-baseline configurations.

Sensor Fusion for Perception:

• Develop sensor fusion pipelines combining EO/IR (thermal/FLIR), depth, and RF data to produce robust environment representations across day/night and degraded conditions.
• Implement probabilistic fusion approaches (complementary filters, tightly-coupled vision-IMU) to maintain perception reliability under low light, motion blur, and occlusion.

Integration & Validation:

• Integrate perception pipelines with ROS2-based software stacks and companion computers (Jetson, RB5).
• Benchmark algorithms in simulation (Gazebo, MATLAB/Simulink) and HIL/HITL testbeds; validate through field trials.
• Measure and report KPIs (detection accuracy, latency, false-positive rate, robustness under environmental variation) and iterate on improvements.
• Document algorithm designs, performance characteristics, and integration interfaces; support certification efforts.

Experience & Skills

Required Qualifications & Experience:

• Master's or PhD in Computer Science, Electrical Engineering, Robotics, or related field.
• 3+ years developing computer vision algorithms for real-world autonomous systems (UAVs, robotics, or autonomous vehicles).
• Proven expertise in object detection, tracking, feature extraction, visual odometry, and camera calibration.
• Proficiency in C++ and C (OpenCV, Eigen) and Python (NumPy, SciPy, scikit-learn, scikit-image).
• Hands-on experience with detection and segmentation frameworks (YOLOv8+, Detectron2, MMDetection, Mask R-CNN, U-Net).
• Advanced use of PyTorch, TensorFlow/Keras, and ONNX for model training and export.
• Experience deploying and optimizing models on embedded platforms (Jetson, Qualcomm RB5, ARM).
• Experience with EO/IR sensor stacks: thermal (FLIR/uncooled) camera integration, NUC calibration, and thermal-domain object detection.
• Strong mathematical foundation in linear algebra, probability theory, and optimization.
• Solid version control practices (Git/GitLab) and dataset/model management (DVC or equivalent).
• English: Upper Intermediate or higher.
• Good communication skills and ability to cooperate with adjacent engineering teams.
• Free criminal record

Preferred Qualifications Experience:

• Proficiency in ROS2: node development, sensor integration, message pipelines.
• Experience with edge AI optimization: quantization, pruning, model compression for TensorRT, ONNX Runtime, OpenVINO.
• Familiarity with annotation platforms (Label Studio, Labelbox) and experiment tracking tools (MLflow,Neptune.ai, TensorBoard).
• Knowledge of tightly-coupled vision-IMU fusion (VINS-Mono, Kimera) for perception-side pose estimation.
• Familiarity with safety-critical standards: DO-178C (avionics software), MISRA C/C++ coding guidelines, STANAG 4671 (UAV airworthiness) as applied to perception pipelines used in flight-critical decisions.
• Track record of publication at CVPR, ICRA, IROS, or IEEE Transactions, or contributions to open-source CV repositories.

Eligibility Requirement:

Candidates must be nationals of a NATO member country or one of the following NATO Indo‑Pacific partner countries: Australia, Japan, South Korea, New Zealand, or Ukraine.

If this sounds like an opportunity that interests you, please send us your application. We look forward to hearing from you!