自动标注算法工程师(Tracking方向)_XC at Bosch Group
Shanghai, Shanghai, China -
Full Time


Start Date

Immediate

Expiry Date

17 Feb, 26

Salary

0.0

Posted On

19 Nov, 25

Experience

2 year(s) or above

Remote Job

Yes

Telecommute

Yes

Sponsor Visa

No

Skills

Computer Vision, Deep Learning, Temporal Modeling, Tracking Algorithms, ID Association, Occlusion Recovery, Trajectory Interpolation, Consistency Maintenance, ReID, Motion Modeling, Track Fusion, Auto-QA, Anomaly Detection, Inference Optimization, Data Throughput, Collaboration

Industry

Software Development

Description
Company Description Do you want beneficial technologies being shaped by your ideas? Whether in the areas of mobility solutions, consumer goods, industrial technology or energy and building technology - with us, you will have the chance to improve quality of life all across the globe. Welcome to Bosch. Job Description 构建面向自动驾驶的 智能自动标注系统,通过融合检测、分割与多目标跟踪(MOT)算法,实现视频与点云数据的时序级自动标注与一致性维护。聚焦于 Tracking算法的设计与工程落地,目标是提升自动标注的时序连贯性、精度与自动化水平,为感知模型训练提供高质量结构化数据。 主要职责 负责视频及点云数据的 多目标跟踪算法(MOT / MOTS / 3D-MOT) 研发与优化; 设计 跨帧 ID 关联、遮挡恢复、轨迹插补、一致性维护 等自动标注核心算法; 研究 ReID、运动建模(Motion Model)、轨迹融合(Track Fusion) 等方法,提升标注稳定性与连续性; 将 Tracking 模块与 Detection / Segmentation 模型输出 融合,形成可闭环自动标注方案; 构建 Auto-QA(自动质检)与异常检测 模块,实现标注一致性与质量可控; 优化推理性能与数据吞吐,支持高并发生产部署; 与感知、仿真、数据挖掘团队协作,推动自动标注系统的算法标准化与生产落地。 Qualifications 硕士及以上学历,计算机科学、人工智能、自动化、电子信息等相关专业; 扎实的 计算机视觉 / 深度学习 / 时序建模 理论基础; 熟悉主流 Tracking 算法(DeepSORT、ByteTrack、BoT-SORT、OC-SORT、CenterTrack、FairMOT、TrackEval 等); 理解 Tracking 与 Detection / Segmentation 联动机制,具备完整 pipeline 实践经验; 良好的工程实现能力与代码规范。 加分项(Highly Preferred): 熟悉 3D Tracking / Multi-Sensor Fusion(Camera + LiDAR + Radar); 熟悉主流检测与分割模型(YOLO, DETR, Mask2Former, Segment Anything 等); 具备 自动标注产线 或 数据生产系统 实际参与经验; 掌握模型推理加速、轻量化、多GPU并行部署; 有学术发表(CVPR / ICCV / ECCV / NeurIPS)或开源项目贡献经验。
Responsibilities
Develop and optimize multi-object tracking algorithms for video and point cloud data. Focus on the design and engineering implementation of tracking algorithms to enhance the quality of structured data for perception model training.
Loading...