Product Manager, Agent Harness & Modelling
Remote • Toronto • FullTime
Posted 4mo ago
About the job
Cohere is seeking an Agent Harness Product Manager to own the execution layer that makes North agents reliable, capable, and production-ready. This role sits at the intersection of agent loop and execution, context engineering, and model-scaffolding co-evolution. You will define how North agents plan and act across long, multi-step workflows, ensuring the execution environment is robust for demanding enterprise tasks. You are expected to engage at the implementation level, contributing to architecture decisions alongside engineering. You will also own how agents manage the context window as a deliberately controlled resource, including progressive disclosure of tools, context compaction, and offloading large payloads. Furthermore, you will own the feedback loop between North's harness and the Modeling Team, ensuring harness design decisions are validated by Modeling and that the model evolves with the harness.
Responsibilities
- Define and own the roadmap for North's agent harness, including the agent loop, context engineering layer, tool orchestration, sandbox execution, and sub-agent delegation.
- Serve as the primary interface between North engineering and Cohere's Modeling team, ensuring new harness capabilities are validated before being built.
- Own North's agentic evaluation framework, ensuring compatibility with harness and modeling infrastructure, and serving as a bridge between product and research.
- Engage enterprise customers to identify agentic failures and translate findings into product and model requirements.
- Stay current with the agent ecosystem and drive adoption decisions aligned with emerging standards.
Requirements
- 5+ years of product management experience in agentic AI systems, developer infrastructure, or applied ML products.
- Deep understanding of modern LLM agent architectures (multi-agent systems, tool-augmented reasoning, memory, retrieval, programmatic orchestration, RAG, long-horizon execution).
- Strong grasp of agentic evaluation design, including measuring task completion, failure recovery, and diagnosing model vs. scaffolding gaps.
- Technical depth to contribute to architecture decisions at the implementation level (design docs, async execution, sandboxed environments, filesystem design).
- Ability to fluently discuss ML research and engineering architecture.
- Track record of shipping platform-layer products with demonstrated impact on reliability, performance, or capability.
Benefits
- Competitive compensation
- Equity options
- Opportunities for professional development
- Hybrid work model