Responsible for the research, development, and engineering implementation of BEV (Bird's Eye View) perception algorithms in autonomous driving systems, including but not limited to tasks such as 2D/3D Object Detection, Semantic Segmentation, Depth Estimation, and Multi-Object Tracking.
Design and optimize BEV feature representations and perception models based on camera, LiDAR, or multi-sensor fusion data to enhance environment understanding in complex scenarios.
Research and implement lightweight deployment strategies (e.g. model compression and inference acceleration) to meet real-time and onboard compute constraints.
Participate in HD map and online mapping development, integrating BEV perception outputs for dynamic environment modeling and cooperative path planning.
Track and explore cutting-edge research (e.g. OCC, NeRF, diffusion models), driving algorithm innovation and real-world application.

Education: Master's degree or above in Computer Science, Artificial Intelligence, Robotics, or related fields.
Core Skills:
- Strong experience in BEV perception algorithm development; familiarity with mainstream models such as DETR3D, BEVFormer, PETRv2, etc.
- Proficient in PyTorch or TensorFlow, with solid C/C++/Python programming skills and experience in model deployment and optimization.
- Good understanding of object detection (e.g. YOLO, CenterNet), multimodal sensor fusion, and 3D reconstruction techniques.
Preferred Qualifications: Experience with autonomous driving simulation platforms (e.g. CARLA, AirSim) or onboard deployment; publications in top-tier conferences such as CVPR, ICRA, or IROS; familiarity with model lightweight techniques including knowledge distillation, model pruning, and quantization.

Perception Algorithm Engineer