Lead Agent Harness PM for Enterprise AI Agents

Cohere; New York (NY); Hybrid; USD 140,000 - 190,000; Full time

Benefits offered by this job

Lunch stipend

Health and dental benefits

RRSP matching

Parental leave

Enrichment benefits

6 weeks vacation

Travel budget for offices

Home office stipend

Job summary

Cohere is seeking an Agent Harness Product Manager to own the execution layer for North, enabling reliable, production-ready agents. You’ll shape the agent loop, context handling, and tool orchestration while working closely with engineering and modeling teams.

You will engage with enterprise customers, validate capabilities before build, and align evaluation with training infrastructure. The role blends ML research conversations with practical architecture decisions within a hybrid work setup.

Qualifications

• 5+ years of product management experience in agentic AI systems, developer infrastructure, or applied ML products.
• Deep understanding of modern LLM agent architectures and task orchestration.
• Strong ability to design agentic evaluation and measure long-horizon reliability.
• Technical depth to contribute to architecture decisions at implementation level.
• Fluent in bridging ML research discussions with engineering design talks.
• Proven track record shipping platform-layer products with impact on reliability or performance.

Responsibilities

• Define and own the roadmap for North's agent harness including loop, context, and orchestration.
• Interface between North engineering and Cohere's Modeling team to validate capabilities.
• Own North's evaluation framework aligning with training infrastructure and product needs.
• Engage enterprise customers to surface failures and translate findings into requirements.
• Monitor open-source and enterprise agent ecosystems to drive adoption.

Skills

Agent loop ownership

Context engineering

Model-scaffolding co-evolution

Job description

Cohere is seeking an Agent Harness Product Manager to own the execution layer for North, enabling reliable, production-ready agents. You’ll shape the agent loop, context handling, and tool orchestration while working closely with engineering and modeling teams.

You will engage with enterprise customers, validate capabilities before build, and align evaluation with training infrastructure. The role blends ML research conversations with practical architecture decisions within a hybrid work setup.