UK - Perception & Vision Systems Engineer -… at MARSS · Bristol

Job Title: Perception & Autonomous Systems Engineer

Reports to: Head of Engineering

Location: Bristol

Type: Full time - permanent

Role Overview

We are seeking a specialist Perception & Vision Systems Engineer to lead the development of high-performance imaging pipelines for our interceptor drone's seeker. This position focuses on the execution and optimisation of vision systems in real-time, high-stakes environments.

The role centers on the architecture of low-latency video pipelines—from MIPI/IP camera drivers through to CUDA-accelerated object detection and tracking. You will be responsible for ensuring our seeker can process high-resolution, high-FPS thermal and optical data, maintaining a stable and accurate state estimate that our guidance systems can rely on.

Key Responsibilities

• Vision Pipeline Architecture & Optimisation
• High-Throughput Processing: Design and optimise pipelines to handle large data packets (high-resolution video frames) at maximum FPS with minimal latency.
• Hardware Acceleration: Implement and tune vision algorithms using OpenCV, CUDA, and TensorRT to leverage the full power of NVIDIA Jetson edge hardware.
• ​GStreamer Experience: Build, debug, and maintain GStreamer pipelines for encoding, decoding, and frame manipulation across multiple camera modalities.
• ​Algorithm Development: Extend existing object detection and tracking systems, utilizing both deep learning and classical methods (e.g., ORB, SIFT, SURF) for robust target locking.

Sensor Integration & Hardware

• Camera Systems: Lead the integration of high-speed MIPI sensors, IP cameras, and
• Thermal (LWIR/MWIR) imaging systems.
• ​Driver Development: Troubleshoot and optimize camera drivers and V4L2 interfaces to ensure stable data acquisition under flight conditions.
• Coordinate Geometry: Manage spatial transformations and coordinate systems to bridge the gap between 2D image-space detections and 3D world-space tracking.

System Integration & Simulation

• ​ROS2 Integration: Maintain the perception-to-system interface via ROS2, ensuring vision outputs are efficiently consumed by the wider autonomous stack.
• Synthetic Data & Simulation: Utilize Unreal Engine, Unity, Gazebo or Other simulators to generate synthetic datasets and validate perception performance in virtual "edge case" scenarios.
• Testing & Validation: Implement rigorous unit and functional testing for perception modules, focusing on deterministic performance and edge-case reliability. Support field testing.

Required Experience

Industry Experience

• ​3+ years in a specialised Vision Engineering or Perception role (UAV, Autonomous
• Vehicles, or Defence).
• ​5–10 years professional engineering experience in C++ and Python, with a focus on
• real-time data processing.

Technical Expertise

• ​Edge Computing: Deep experience with the NVIDIA Jetson ecosystem (Orin/Xavier) and
• CUDA acceleration.
• Video Infrastructure: Expert-level knowledge of GStreamer, video codecs (H.264/H.265), and high-bandwidth image handling.
• ​Computer Vision: Proficiency in OpenCV and the deployment of models via TensorRT.
• ​Spatial Math: Comfortable with 3D geometry, coordinate transformations, and sensor calibration workflows.
• Embedded Linux: Comfortable working in Linux environments, including driver-level troubleshooting and system-level performance tuning.
• Git and Agile Methodologies

Preferred Skills

• Experience with Thermal/Infrared imaging and multi-spectral sensor fusion.
• Familiarity with classical feature extraction and matching (SURF, SIFT, ORB) for GPS-denied navigation or tracking.
• Hands-on experience with synthetic data generation pipelines (AirSim, Gazebo, or custom Unreal/Unity plugins).
• Understanding of IP-based vision protocols (RTSP/UDP) and low-level MIPI CSI-2 specifications.
• ​Object oriented programming.
• Full Driving License - willingness to travel to support testing and demos

Attributes & Mindset

• ​Data-Centric: You treat every frame as a critical packet of information where latency is the primary enemy.
• ​Hardware-Aware: You don't just write code; you write code that understands the constraints and capabilities of the silicon it runs on.
• ​Iterative Prototyper: Comfortable in fast-paced R&D, moving quickly from a Python proof-of-concept to a highly optimised C++ production node.
• Rigorous Tester: Dedicated to functional safety, ensuring the seeker performs reliably in diverse lighting and thermal conditions.
• ​Comfortable working with multiple engineering disciplines in a small tight-knit team.
• Adaptable & Resilient: Stay calm and focused under pressure, maintaining a positive attitude that motivates others, adapting approach as challenges arise.

We believe in the value of working together in person and expect team members to be office-based for the majority of the working week. At the same time, we take a flexible pragmatic approach to working patterns, including start and finish times and remote working, to support a healthy work-life balance while meeting the needs of the team and project.

We are committed to building an inclusive and diverse team where everyone is treated with respect and has the opportunity to contribute and succeed. We welcome applications from people of all backgrounds and are focused on creating a working environment where different perspectives, experiences and ideas are valued