Qualifications
Qualifications:Bachelor's degree or higher in Computer Science, Automation, Electronics, Image Processing, or related fields. Understanding of the overall architecture of autonomous driving perception systems, particularly the engineering role of image pipelines in algorithm preprocessing. Proficient in basic image processing concepts (color space, geometric transformations, interpolation, distortion correction, etc.), with a focus on engineering implementation. Strong programming skills in C / C++ and experience in maintaining complex engineering code. Familiarity with at least one mainstream autonomous driving computing platform's imaging or acceleration frameworks: NVIDIA (CUDA / GPU Image Pipeline / VIC, etc.)Qualcomm (DSP / GPU / FastCV / HVX, etc.)Experience with heterogeneous computing and understanding of engineering constraints such as throughput, latency, bandwidth, and memory usage in algorithm pipelines. Strong Plus:Experience in engineering perception algorithm pipelines (e.g., object detection, lane detection, OCC, automated parking, etc.). Experience with image processing or perception preprocessing pipeline development on Horizon chips. Familiarity with multi-camera systems (front view / surround view / fisheye) and their usage in perception algorithms. Experience in mass production projects and understanding the complete process from algorithm requirements to engineering delivery. Familiarity with cross-platform perception system development under Linux / QNX.
About the job
This position focuses on the development of image processing pipelines within autonomous driving perception systems.
The role involves delivering stable, controllable, and high-performance image data inputs for perception algorithms (such as Detection, Lane, OCC, Parking, etc.).
Responsibilities
- Design and develop the image processing pipeline in the perception system to meet the unified input requirements of multiple perception tasks.
- Develop and maintain image processing modules that support perception algorithms, including but not limited to:
- Image pre-processing / format and resolution conversion
- Geometric correction, coordinate alignment, time synchronization
- Cropping, stitching, and normalization tailored to algorithm needs
- Develop and optimize heterogeneous compute kernels (CPU / GPU / DSP / NPU) based on performance and latency requirements of perception algorithms.
- Collaborate closely with the perception algorithm team to define input specifications, data format standards, and interface designs.
- Ensure the stable operation and performance optimization of perception algorithms across various chip platforms (such as NVIDIA, Qualcomm).
- Oversee the stability assurance, problem diagnosis, and long-term maintenance of production-level image pipelines.
About Bosch China Innovation and Software Development Campus
Bosch China Innovation and Software Development CampusAt Bosch's Innovation and Software Development Campus, we focus on the electrification, automation, connectivity, and personalization of transportation solutions, providing innovative technologies and cutting-edge solutions for the future and accelerating the realization of technology strategies tailored to the Chinese market. Our main development directions include intelligent connected vehicles, smart cockpits, advanced driving assistance, vehicle-road-cloud collaboration, Industry 4.0, artificial intelligence big data, smart home, and embedded software services. The Bosch China Innovation and Software Development Campus empowers customers through software, achieving innovative research and development results in fields such as automotive assisted driving, hydrogen fuel cells, electric drive axles for heavy trucks, and multi-functional controllers.
