Autonomous Driving Data & Platform Engineer
Multiple Locations4d ago
Free account needed to apply or save. It takes a minute, and we'll bring you right back to this job.
Skills
About the role
China is Mercedes-Benz’s largest passenger car market globally. At Mercedes-Benz R&D China, we deliver China-fit ADAS/AD solutions through cutting-edge technologies, and we are seeking talented professionals who share this passion.
Responsible for building and evolving the AD data closed-loop toolchain that supports Mercedes-Benz China ADAS/AD mass-production programs and empowers algorithm and validation teams across R&D China. The role covers three pillars — scenario management and evaluation infrastructure, ADAS simulation and Component SiL support, and AI agent tooling — and grows alongside the industry shift toward end-to-end, data-driven autonomous driving. The ideal candidate is a self-driven engineer with a product mindset and strong interest in AI/LLM, who takes ownership from idea to adoption.
Key Responsibilities
Scenario Management & Evaluation Infrastructure
- Scenario Tree: Design and maintain the scenario tree to support hierarchical organization, coverage tracking, and trend analysis.
- DScenario Analytics: Develop scenario statistics and visualization to improve interpretability and management efficiency.
- Evaluation Pipeline: Build evaluation infrastructure, including metric frameworks, automated reporting, and regression analysis, to assess algorithm performance and data quality and to support closed-loop evaluation of end-to-end models.
ADAS Simulation & Component SiL Support
- Simulation Support: Assist in setting up and maintaining scenario-based and log re-simulation environments
- Component SiL Integration: Help integrate ADAS components into the SiL environment and coordinate with algorithm teams on interface adaptation and issue troubleshooting
- Test Automation: Contribute to automated simulation execution and regression testing, linking simulation results with the scenario library and evaluation pipeline
AI Agent & Intelligent Tooling for Autonomous Driving
- AI Agent Tools: Design, develop, and iterate on AI agents for AD R&D workflows, such as data quality assessment, evaluation metric development, and report generation
- LLM Integration: Drive LLM integration into the closed-loop toolchain to automate data production, evaluation, and operations end-to-end
- AI-Driven Innovation: Propose and prototype AI-driven approaches for scenario mining, badcase analysis, and data selection to boost R&D efficiency
China is Mercedes-Benz’s largest passenger car market globally. At Mercedes-Benz R&D China, we deliver China-fit ADAS/AD solutions through cutting-edge technologies, and we are seeking talented professionals who share this passion.
Responsible for building and evolving the AD data closed-loop toolchain that supports Mercedes-Benz China ADAS/AD mass-production programs and empowers algorithm and validation teams across R&D China. The role covers three pillars — scenario management and evaluation infrastructure, ADAS simulation and Component SiL support, and AI agent tooling — and grows alongside the industry shift toward end-to-end, data-driven autonomous driving. The ideal candidate is a self-driven engineer with a product mindset and strong interest in AI/LLM, who takes ownership from idea to adoption.
Key Responsibilities
Scenario Management & Evaluation Infrastructure
- Scenario Tree: Design and maintain the scenario tree to support hierarchical organization, coverage tracking, and trend analysis.
- DScenario Analytics: Develop scenario statistics and visualization to improve interpretability and management efficiency.
- Evaluation Pipeline: Build evaluation infrastructure, including metric frameworks, automated reporting, and regression analysis, to assess algorithm performance and data quality and to support closed-loop evaluation of end-to-end models.
ADAS Simulation & Component SiL Support
- Simulation Support: Assist in setting up and maintaining scenario-based and log re-simulation environments
- Component SiL Integration: Help integrate ADAS components into the SiL environment and coordinate with algorithm teams on interface adaptation and issue troubleshooting
- Test Automation: Contribute to automated simulation execution and regression testing, linking simulation results with the scenario library and evaluation pipeline
AI Agent & Intelligent Tooling for Autonomous Driving
- AI Agent Tools: Design, develop, and iterate on AI agents for AD R&D workflows, such as data quality assessment, evaluation metric development, and report generation
- LLM Integration: Drive LLM integration into the closed-loop toolchain to automate data production, evaluation, and operations end-to-end
- AI-Driven Innovation: Propose and prototype AI-driven approaches for scenario mining, badcase analysis, and data selection to boost R&D efficiency
Education
Master degree in field such as Computer Science, Electrical Engineering, Robotics, Automotive Engineering or equivalent
Experience
- More than 5 years of software development experience, preferably in autonomous driving, vehicle software, or simulation domains
- Minimum 2 years of hands-on experience in ADAS toolchain or AI application development
Technical Requirements
Python Development: Proficient in Python backend and data processing, including API design, databases (SQL/MongoDB), caching (Redis), and pandas
- AD Domain Knowledge: Familiar with AD development workflows such as data closed-loop, scenario libraries, evaluation metrics, and simulation/SiL testing
- C++ Proficiency: Able to read and debug C++ code for ADAS component integration in SiL environments
- AI/LLM Application: Hands-on experience building LLM-based applications (e.g., RAG, tool calling, MCP, or agent frameworks such as LangGraph)
- Language: Working proficiency in English
Mindset & Soft Skills
- Ownership & Drive: Proactively identify pain points and drive solutions from design to launch without waiting for detailed specifications
- Product Thinking: Treat internal tools as products — understand user needs, prioritize by impact, and iterate based on feedback