Intern_自动驾驶数据产线Agent开发实习生

[6 months, 4-5 days/week]

Name of Project: AD Data Pipeline Agentic Systems Intern - 自动驾驶数据产线Agent开发实习生

Objectives:

Mercedes-Benz is developing scalable data pipelines to support autonomous driving. This project introduces LLM-based agents and intelligent automation across the data lifecycle, including data ingestion, processing, annotation, quality validation, and delivery. You will help design and implement agent workflows, internal tools, and system integrations to improve efficiency, reduce manual work, and enhance process reliability.

Main Tasks:

- Design and implement LLM-based agentic workflows for data ingestion, cleaning, annotation dispatching, quality assurance, and other repetitive data-pipeline tasks. Deliver working prototypes and reusable workflow components (30%)

- Build and maintain internal toolchains, CLI utilities, and automation services to improve developer&operator productivity (25%)

- Integrate agents with existing data platforms, labeling systems, CI/CD pipelines, and cloud services through well-defined tools and interfaces (20%)

- Implement robust orchestration logic, incl. error handling, retry mechanisms, human-in-the-loop escalation, logging, monitoring, and execution traceability (15%)

- Evaluate and benchmark agent performance, iterate on prompts and tool definitions, and document technical designs, evaluation results, and development best practices. (10%)

Qualification:

Bachelor/Master student in Computer Science, Software Engineering, AI or related fields

Strong Python programming skills

Experience with LLM APIs, prompt engineering, tool calling, agent workflows and frameworks (LangChain, LangGraph, CrewAI, AutoGen, etc.)

Familiar with Linux, Git, CLI tools and software engineering best practices

Strong analytical, problem-solving and self-learning abilities

Preferred

Familiar with TypeScript, Go, Rust, data orchestration tools (Airflow, Prefect, Temporal, Dagster), containerization and cloud technologies (Docker, Kubernetes, AWS, GCP, Azure)

Experience building internal tools, automation workflows and integrations with data platforms or CI/CD systems

Bonus

Experience with autonomous driving data, MCP servers, multi-agent systems, LLM evaluation frameworks, or related open-source projects

Language:

Chinese: Proficient

English: Conversational