[6 months, 4-5 days/week] Objectives: Mercedes-Benz are developing high-precision AD/ADAS systems with a strong emphasis on sensor perception quality. To enable large-scale deployment of reliable autonomous driving functions, we are building robust online calibration pipelines that ensure accurate multi-sensor (LiDAR-Camera-IMU) alignment under real-world driving conditions. The intern will work as an Online Calibration Algorithm Developer, focusing on the development and optimization of targetless calibration frameworks. Working closely with the team leader and senior algorithm engineers, the intern will contribute to the design, implementation, and deployment of production-ready calibration pipelines, ensuring data quality and system robustness across large-scale fleet. Main Tasks: - Online calibration dataset processing from real world driving scenarios. (20%) - Online calibration algorithm research, development and software pipeline development . (60%) - Design and implement evaluation metrics to assess calibration accuracy, robustness, and runtime performance. (20%) Learning Opportunities: 1. Develop online calibration algorithms together with MB algorithm team, gaining experience with production-level calibration frameworks and sota optimization techniques. 2. Work with large-scale fleet data to understand real-world sensor degradation, extrinsic drift, and dynamic calibration challenges in autonomous driving systems. 3. Opportunities to work on the project as a pure developer with Mercedes-Benz algorithm and software engineering team. 4. Explore and validate new ideas on the Mercedes-Benz large scale autonomous driving dataset. Key Qualifications: 1. Bachelor’s degree or above in Computer Science, Mechanical Engineering, Automation, Robotics, or a related field; 2. Solid education background on computer science and software engineering. 3. Familiar with C++ and Python, with experience in Linux development environment and toolchains (Git, CMake, etc.). 4. Foundation in linear algebra and 3D geometry; familiar with numerical optimization methods (e.g., Levenberg-Marquardt, Gauss-Newton); understands the principles of hand-eye calibration or PnP problem solving.. 5. Hands-on experience with point cloud libraries (PCL or Open3D), including filtering, feature extraction, and point cloud registration (ICP/NDT). Language: Chinese: Proficient English: Proficient Education: Degree: Bachelor or above Major: Computer Science