智能驾驶端到端算法研究科学家_CR

Location: Shanghai, Shanghai, cn

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

• 基于摄像头、雷达和激光雷达，研发自动驾驶相关的端到端算法，例如用于驾驶决策的两阶段与单阶段模型。
• 参与高级驾驶辅助系统端到端流水线的车载设计与部署，实现上下游模块间的协同交互。
• 与博世中国及全球业务部门合作，进行需求分析、技术转移，并评估新概念的市场潜力。
• 跟踪、识别并探索中国本土的新趋势与新兴技术，分析其对博世研发战略的影响，发掘关键合作伙伴。
• Research and development of E2E related algorithms for autonomous driving, based on camera, radar and lidar, such as two-stage and one-stage models for driving.
• Participate in the design and deployment of ADAS E2E pipeline on vehicle, with interaction between upstream and downstream modules.
• Cooperation with Bosch business units in China and worldwide for requirement analysis, technology transfer, and to evaluate market attractiveness of new concepts.
• Scouting, identification, and tracking of new trends and emerging technologies in China, deriving impacts on Bosch research strategy, and identifying key partners.

Qualifications

• 拥有计算机科学、电气工程、机电一体化、机械工程或相关领域的优秀硕士学位，博士学位优先。
• 具备扎实的端到端模型领域知识，尤其在单感知、多阶段训练策略及基础模型蒸馏方面。
• 熟悉端到端前沿方法，如VAD、PDM-Hybrid、扩散模型、sparsedrive、diff-vla、IRL-VLA、world4drive等。
• 深入理解感知域预训练策略，包括通用地图元素检测任务、动态目标检测与跟踪任务、自监督任务及无监督感知任务。
• 具有基础模型蒸馏策略相关经验。
• 拥有端到端高级驾驶辅助系统研发经验者优先。
• 熟悉现代人工智能框架：Pytorch、OpenMMLab、HR HAT。
• 具有在车载高级驾驶辅助系统计算单元上设计与部署感知模块的经验。
• 熟悉Python或C++编程语言。
• 熟悉Linux操作系统。
• 具备摄像头、雷达、激光雷达等高级驾驶辅助系统传感器相关经验。
• 具有国际交流或合作经验者优先。
• 热爱创新，善于解决问题，具备创业精神。
• 强烈的团队合作意识，能够自主高效地开展工作。
• 优秀的沟通能力，英语流利（口语与书面）。
• Excellent master's degree, ideally PhD, in the field of computer science, electrical engineering, mechatronics, and mechanical engineering or related.
• Strong knowledge in E2E model domain, especially one-perception, multi-stage training strategy and foundation model distillation.
• Strong knowledge in E2E SOTA methods, such as VAD, PDM-Hybrid, Diffusion Models, sparsedrive, diff-vla, IRL-VLA, world4drive.
• Deep understanding on perception space pre-training strategies, such as general map element detection tasks, dynamic object detection & tracking tasks, self-supervised tasks and unsupervised perception tasks.
• Have experience in foundation model distillation strategies.
• Experience on E2E ADAS research & development is a strong plus.
• Familiarity with modern AI frameworks: Pytorch, OpenMMLab, HR HAT.
• Experience with designing and deploying perception on onboard ADAS computation unit.
• Familiarity with Python, or C++.
• Familiarity with Linux.
• Experience in ADAS sensors, such as cameras, radars, and LiDARs.
• International experience is a plus.
• Passion for innovation, problem-solving & entrepreneurial mindset.
• Strong team player with an autonomous and efficient working style.
• Excellent communication skills, fluent in English (oral and written).