You’ll be the glue between sensors, software, and hardware—owning system integration of LiDAR + camera stacks on a ROS2 platform (Jetson Orin) and taking end-to-end responsibility for validation & test (SIL/HIL, on-rig, on-road).

Tasks

• System integration (LiDAR + cameras): Bring up, configure, and tune sensors (e.g., Ouster/Hesai + GMSL2 cameras); manage drivers, time sync (PTP/IEEE-1588), triggering, exposure, and bandwidth.

• ROS2 on Orin: Package perception nodes (C++/Python), deploy with Docker on NVIDIA Jetson Orin; optimize with Isaac ROS/TensorRT where relevant.

• Calibration & perception pipeline: Own camera intrinsics/extrinsics and LiDAR–camera calibration (e.g., Kalibr, OpenCV, PCL); verify point-cloud ↔ image alignment and perception KPIs.

• Data + logging: Define rosbag2 logging strategies, dataset/versioning hygiene, and replay workflows (SIL/HIL) for deterministic debugging and regression.

• Test ownership: Define the test strategy (component → system), write test plans, create automated launch/integration tests (gtest/pytest/ROS2 launch), and drive CI (build, hardware-smoke, regression).

• Reliability in the loop: Proactively chase down flaky drivers, clock drift, thermal throttling, and real-time constraints; contribute to HW BOM choices (cables, hubs, PPS/GPS time).

• Collaboration: Work across Perception, Firmware, and Platform; document architecture and test results; support on-rig bring-up and field testing.

Requirements

Must have:

• 4–8+ years in robotics/AV/vision system integration (not just pure research).

• Strong ROS2 (rclcpp, launch, DDS like Cyclone/FastDDS), C++14/17 (some Python), CMake/Bazel, Git, Docker.

• Hands-on with LiDAR (Ouster/Hesai/Velodyne) & industrial cameras (GMSL2/CSI/USB3; FLIR/Basler/etc.), time sync (PTP/NTP/PPS) and calibration tooling.

• Comfortable on NVIDIA Jetson Orin (JetPack, Nsight, tegrastats), performance/latency profiling, TensorRT/Isaac ROS (plus).

• Proven test ownership: designing automated integration/system tests, SIL/HIL, hardware smoke tests, and CI pipelines.

• Linux/Ubuntu (20.04/22.04), networking (Ethernet, VLANs), basic CAN.

Nice-to-have:

• Isaac ROS GEMs / Isaac Sim / Ignition Gazebo; CUDA/VPI.

• ASPICE/ISO-26262 exposure (mindset for safety), MISRA C++ familiarity.

• OpenCV, PCL, Open3D; exposure to label tools (CVAT/Label Studio).

• German language (nice) — English required.

Join us
If you love building reliable perception systems and want to see your work run on real hardware, in the field, and on the road - we’d love to meet you.