混元AI Agent Harness Engineer（北京/深圳））
Core
Design and implement tracing & observability systems, quality evaluation pipelines, and debugging tools for AI Agents; engineer the collaboration between large models and agents.
Role type
Senior IC AI Agent Harness Engineer
Builds
Tracing/observability systems, automated eval pipelines, A/B testing frameworks, regression detection tools, and core support platforms for data and evaluation.
Domain
Artificial Intelligence / Large Language Models / Agent Systems
Deliverable
production ML models
Required skills
Agentic coding, full-stack engineering, system design, LLM integration, Agent framework development, prompt engineering, code quality and architecture
Preferred skills
Experience with Cursor/Claude Code/Codex, deep understanding of AI Agent workflows, ability to work closely with research teams
Technologies
LLM, Agent frameworks, prompt engineering, automated evaluation pipelines, A/B testing, regression detection
Responsibilities
Design and implement tracing & observability systems for Agent execution; Build automated evaluation pipelines, A/B testing, and regression detection; Develop debugging tools for Agents; Architect and develop core support platforms for labeling, evaluation, and data pipelines; Collaborate with AI research teams to engineer large model and Agent co-evolution; Explore and practice daily workflows for AI Agents.
