计算机视觉算法工程师_BCSC

Bosch China Innovation and Software Development Campus

博世创新与软件开发中心

    博世创新软件开发（无锡）有限公司针对交通出行的电动化、自动化、互联化、个性化，提供面向未来的创新技术和前沿解决方案，加速针对中国市场的技术战略的实现和发展。博世软件中心主要发展方向包括智能网联汽车、智能座舱、自动驾驶、车路云协同、工业4.0、人工智能大数据、智能家居、嵌入式软件服务。博世中国创新与软件开发中心以软件为客户赋能，在汽车自动驾驶、氢燃料电池、重卡电驱动桥、多合一的控制器等多个领域取得创新研发成果。

我们正在寻找一位专注于驾驶辅助系统的感知算法工程师。您将负责开发高精度的车灯与交通信号灯检测算法，并部署于地平线征程J6系列芯片。

We are seeking a dedicated Algorithm Engineer focused on ADAS perception. You will be responsible for developing high-precision vehicle light and traffic light detection algorithms and deploying them on the Horizon Robotics Journey J6 series chips.

主要职责 / Key Responsibilities:

• 核心算法研发 / Core Algorithm Development:• 研发基于深度学习的车灯检测（前照灯、尾灯、刹车灯、转向灯）与交通信号灯检测识别算法。

Develop deep learning-based algorithms for vehicle light (headlights, tail lights, brake lights, turn signals) and traffic light detection and recognition.
• 专注解决小目标、远距离及极端光照（如夜间、隧道口）等场景下的技术挑战。

Focus on solving technical challenges in scenarios involving small objects, long distances, and extreme lighting conditions (e.g., night, tunnel entrances).

• 平台部署与优化 / Platform Deployment & Optimization:• 主导模型在地平线J6系列芯片上的部署、调试与性能优化。

Lead the deployment, debugging, and performance optimization of models on Horizon J6 series chips.
• 熟练使用地平线天工开物(OpenExplorer)工具链进行模型转换与量化。

Proficiency in using the Horizon OpenExplorer toolchain for model conversion and quantization.

• 数据闭环 / Data Closed-Loop:• 构建相关数据集，并通过数据驱动策略持续提升模型性能。

Build relevant datasets and continuously improve model performance through data-driven strategies.

• 工程实现 / Engineering Implementation:• 编写高质量、可维护的C++/Python代码，并参与构建自动化测试流程。

Write high-quality, maintainable C++/Python code and participate in building automated testing pipelines.

必备条件 / Required:

• 学历与经验 / Education & Experience:• 计算机、电子工程等相关专业本科及以上学历，具备2年以上计算机视觉算法开发经验。有ADAS/自动驾驶经验者优先。

Bachelor's degree or above in Computer Science, Electronic Engineering, or related fields. 2+ years of experience in computer vision algorithm development. Experience in ADAS/autonomous driving is a plus.

• 编程能力 / Programming Skills:• 精通Python，熟练掌握C++。

Proficient in Python, skilled in C++.

• 技术栈 / Technical Stack:• 精通PyTorch或TensorFlow，熟练使用OpenCV。

Proficient in PyTorch or TensorFlow, familiar with OpenCV.

优先考虑 / Preferred Qualifications:

• 平台经验 / Platform Experience:• 拥有地平线征程系列芯片（J5/J6）模型部署和优化经验，熟悉天工开物工具链。

Hands-on experience with model deployment and optimization on Horizon Journey series chips (J5/J6), familiar with the OpenExplorer toolchain.

• 领域知识 / Domain Knowledge:• 具备车灯/交通灯检测相关项目经验。

Project experience in vehicle lights/traffic light detection.

• 优化能力 / Optimization Skills:• 具备模型量化（INT8）、剪枝等优化实战经验。

Practical experience in model optimization techniques such as quantization (INT8) and pruning.