Senior Software Engineer - Perception

Role Overview

This position involves comprehensive involvement in the perception technology stack, including sensor calibration, multi-sensor fusion, and development of vision models deployed onboard robotic platforms and within data processing pipelines. As a key contributor in a compact team, the role will adapt to your expertise with opportunities to assume full technical leadership over the perception stack.

Key Responsibilities

• Perform sensor calibration tasks across multiple modalities such as depth sensors, 3D LiDAR, RGB cameras, Time-of-Flight, and sonar, including performing in-field calibration procedures.
• Manage multi-sensor data fusion, maintain the stability of pointcloud processing pipelines, and ensure CPU usage remains within allocated limits.
• Develop and optimize onboard perception models for object detection and segmentation, including generating zone proposals within mapped environments.
• Work on offboard vision models focused on automatic labeling, data mining, and synthetic data creation.
• Engage in research and prototyping with emerging vision model technologies, including vision-language integration, monocular depth estimation, and 3D reconstruction for simulation and synthetic data generation.
• Maintain the data processing pipeline and oversee the lifecycle management of datasets and models, including version control, experiment tracking, and secure over-the-air updates for models.
• Implement automated testing frameworks such as benchmarks and software-in-the-loop regression tests to serve as release acceptance criteria.

Eligibility and Qualifications

• Bachelor’s or Master’s degree in Computer Science, Electrical Engineering, Robotics, or a closely related discipline.
• At least 4 years’ experience in perception or computer vision roles within real-world product environments.
• Proficiency in Python and PyTorch, alongside C++ for embedded/onboard programming.
• Hands-on expertise covering both sensor data processing (calibration, pointcloud processing, depth measurement, LiDAR) and vision model training on self-collected datasets; expertise in only one area is insufficient.
• Experience deploying models to embedded compute platforms using technologies such as OpenVINO, TensorRT, ONNX, RKNN, or Hailo.
• ROS 2 experience is highly regarded and considered a strong advantage.