As a Perception Engineer in our company, your role will revolve around the design, development, and optimization of advanced real-time perception components for our autonomous robotics systems, with an emphasis on handling point cloud data from lidars as well as image processing. You will also be tasked with the integration of the perception software into these systems, as well as with testing and verification of the algorithms through both simulation and field experiments.
Responsibilities
- Development, optimization, and maintenance of core classical and deep learning based perception algorithms for our products, focusing on, but not limited to, LiDAR data processing.
- Integration of perception components into one or more ROS-based systems and maintenance of the interfaces.
- Establishment of testing procedures to ensure the stability, functionality, and performance of the perception components.
- Troubleshooting perception-related technical issues in our systems, proposing and implementing effective solutions to them.
- Continuous enhancement of the performance, efficiency, and scalability of the perception algorithms and components.
- Closely collaborate with other perception team members and project teams, and communicate developments in perception algorithms to them and, as appropriate, to the broader company
Requirements
- Bachelor’s, Master’s degree or higher in Computer Science, Robotics, or a related field. Candidates without a degree must have significant software development experience.
- Proficiency with C++ OR Python.
- 2+ years experience as a Software Engineer developing classical perception algorithms, particularly for point cloud data processing, using libraries such as Open3D and PCL, and image processing using libraries such as OpenCV.
- Basic experience with Robot Operating System (ROS) and a general understanding of its capabilities.
- Basic experience with deep learning frameworks, ideally PyTorch.
- Experience developing in a Linux environment and working with large codebases, with Git-based version control, and basic knowledge of Docker for containerization.
Nice to haves
While not specifically required, tell us if you have any of the following.
- Experience working in the autonomous vehicle industry; hands-on experience with sensors such as lidars, cameras, IMU, GNSS, etc., as well as CAN bus data is a plus.
- Strong knowledge of best practices in software development.
- Proficiency in C++ AND Python, including scalable library development and deployment with tools such as Poetry, CMake, CPack, and CPM.
- Advanced classical and deep learning based perception knowledge (e.g. visual/lidar odometry, image inpainting, customizing deep learning model architectures, MLOps).
- Advanced knowledge of ROS (e.g. DDS, core-middleware, QOS, node health monitoring).
- Familiarity with Agile and the Scrum methodology.
- Proficiency in the Japanese language.