Job opportunity
Agent Harness Product Manager
Cohere
Own agent runtime architecture and context engineering at Cohere, shaping reliable North workflows for enterprise AI deployments. This is a hybrid role.
As an Agent Harness Product Manager, you will lead North’s execution layer across agent loops, tool orchestration, and sandbox execution. You will connect engineering and Modeling through shared evaluations, architecture decisions, and product requirements. The role centers on making agents reliable, capable, and production-ready for demanding enterprise tasks.
Key Responsibilities:
Define North’s agent harness roadmap and execution capabilities
Coordinate harness validation between engineering and Modeling
Own agentic evaluations across product and research infrastructure
Translate enterprise agent failures into product requirements
Track ecosystem developments and guide adoption decisions
Requirements:
Five-plus years managing agentic AI, infrastructure, or applied ML products
Deep knowledge of LLM agents, RAG, retrieval, and orchestration
Strong understanding of agentic evaluation and failure diagnosis
Technical fluency in architecture, asynchronous execution, and sandboxes
Experience shipping platform products improving reliability or capability
Shape North’s agent runtime, evaluations, and context systems while connecting Cohere’s engineering and Modeling teams.
#J-18808-Ljbffr
Apply on source site As an Agent Harness Product Manager, you will lead North’s execution layer across agent loops, tool orchestration, and sandbox execution. You will connect engineering and Modeling through shared evaluations, architecture decisions, and product requirements. The role centers on making agents reliable, capable, and production-ready for demanding enterprise tasks.
Key Responsibilities:
Define North’s agent harness roadmap and execution capabilities
Coordinate harness validation between engineering and Modeling
Own agentic evaluations across product and research infrastructure
Translate enterprise agent failures into product requirements
Track ecosystem developments and guide adoption decisions
Requirements:
Five-plus years managing agentic AI, infrastructure, or applied ML products
Deep knowledge of LLM agents, RAG, retrieval, and orchestration
Strong understanding of agentic evaluation and failure diagnosis
Technical fluency in architecture, asynchronous execution, and sandboxes
Experience shipping platform products improving reliability or capability
Shape North’s agent runtime, evaluations, and context systems while connecting Cohere’s engineering and Modeling teams.
#J-18808-Ljbffr