上海-Agent Harness 研发工程师(J100779)
Core
Design and develop the core technical architecture for Agent Harness to enable stable LLM capabilities in real-world business scenarios.
Role type
Senior IC machine-learning engineer (AI Agent systems)
Builds
Production-ready Agent Harness systems supporting state management, sandbox environments, and execution constraints for complex tasks.
Domain
Artificial Intelligence / Large Language Models / Agent Systems
Deliverable
production ML models
Required skills
PyTorch, LangChain, AutoGen, ReAct, CoT, Plan-and-Execute, system architecture design, context management, feedback loop design
Preferred skills
Open source contributions to AI Agent projects, early-stage product building experience, top-tier conference publications, technical competition awards
Responsibilities
Design and implement collaboration mechanisms between Agents and humans, analyze and validate Harness designs against model behavior, optimize context handoffs and information compaction, drive product iteration based on user feedback.
