Key Skills

Job Description

Ark exists to protect the people of Earth. We build advanced robotic systems for defence and other demanding real-world environments, bringing together robotics, autonomy, hardware, software, manufacturing and frontline operations. Our teams work across several countries and closely with the people who use our systems in the field. We are entering the next stage of the company's growth. Products must develop faster. Manufacturing must scale. Customer commitments must stay connected to delivery reality. The role The autonomy team’s perception workload is growing rapidly with new platform variants and operational requirements. Working alongside the autonomy lead, you will be a key individual contributor pushing our perception capabilities forward. This is a research-first role with a direct path to deployed systems. You will tackle edge-case problems—such as dynamic battlefield geometries and active camouflage—while operating under strict computational constraints. Expected Outcomes Rapid Deployment: Ship one major technical contribution within your first 2 months. Direct Impact: Contribute to at least one fully deployed perception module on our autonomous hardware. Applied Research: Transition complex theoretical models (like thermal fusion and anomaly detection) into practical, deployable code. Сore responsibilities Advanced Sensor Fusion: Design and implement sensor fusion algorithms integrating RGB, LiDAR, Radar, and Thermal/IR imaging. Adversarial & Dynamic Detection: Develop novel algorithms to detect personnel and equipment employing active camouflage, and instantly map complex, fast-changing 3D geometries to distinguish clear routes from blocked terrain. Algorithm Optimization: Train and optimize deep learning models and probabilistic algorithms to run efficiently on edge computing devices (NVIDIA Jetson, FPGA) without sacrificing mission-critical latency. System-Level Integration: Build, test, and iterate on robotic prototypes, ensuring perception modules integrate seamlessly into our ROS2 architecture. Qualifications & Competencies Education: MS or PhD in Robotics, Computer Science, Mathematics, Physics, or a related field requiring deep theoretical math. Experience: 5+ years of industry experience in robotics R&D. Strong publications or an equivalent applied research track record. Technical Stack: Proficiency in Python and C++ within ROS2. Expertise with PyTorch or JAX. Experience optimizing neural networks for embedded hardware (TensorRT, CUDA). Domain Expertise: Mastery of computer vision (3D object detection, depth estimation, semantic segmentation) and multi-modal sensor fusion techniques (Kalman Filters, Particle Filters, Bayesian networks). Defense Context: Experience with military or dual-use technologies, modeling environmental noise, and handling sensor uncertainty in contested environments is preferred.

Core Responsibilities

Design and implement multi-sensor perception and detection algorithms for dynamic, adversarial environments, including fusion of RGB, LiDAR, radar, and thermal/IR data. Optimize models for edge devices and integrate, test, and iterate on perception modules within ROS2-based robotic systems.

Requirements

Requires a master's or doctoral degree in robotics, computer science, mathematics, physics, or a related field, plus at least five years of industry experience in robotics R&D. Candidates should have strong research credentials and expertise in computer vision, sensor fusion, Python and C++ with ROS2, deep learning frameworks, and embedded model optimization; defense or dual-use experience is preferred.