Product Manager, Agent Harness & Modelling

About Cohere

Cohere is a pioneer in secure, enterprise AI, developing advanced foundation models and comprehensive AI products aimed at solving real business challenges. With offices worldwide, including Montreal, Toronto, London, and others, Cohere's diverse and passionate team includes researchers, engineers, and designers focused on advancing AI capabilities and delivering exceptional value to customers.

About North Platform

North is Cohere's agentic AI platform engineered to securely deploy AI agents and automations across enterprise infrastructures. It empowers organizations to optimize workflows, automate repetitive tasks, and derive actionable insights while maintaining strong data privacy and compliance. North integrates state-of-the-art generative and search models with customizable integrations to enhance productivity and innovation.

Role Overview

The Agent Harness Product Manager will lead the development of the execution layer that ensures North's AI agents are dependable, capable, and production-ready. This role intersects three key domains:

• Agent Loop and Execution: Manage the core runtime including tool orchestration, parallel processing, delegation to sub-agents, sandboxed code execution, and failure handling. Define strategies for multi-step workflows and ensure the execution environment supports enterprise-grade demands.
• Context Engineering: Oversee the management of context window resources through controlled tool disclosure, context summarization, payload offloading to persistent storage, and instrumentation to maintain agent orientation during extended actions.
• Model-Scaffolding Co-evolution: Facilitate connectivity between the harness and modeling teams by validating design decisions before implementation, coordinating evaluation frameworks, and aligning model evolution with harness enhancements.

Key Responsibilities

• Develop and own the roadmap for North's agent harness components including agent loop, context engineering, tool orchestration, sandbox execution, and delegation.
• Serve as principal liaison between North’s engineering and Modeling teams, ensuring coordination and validation of new harness features prior to development.
• Manage North's agent evaluation framework, ensuring compatibility with harness and modeling infrastructures and bridging product and research efforts.
• Engage with enterprise clients to identify agent failure modes and translate insights into product and modeling requirements.
• Collaborate with modeling experts to set objectives maximizing model impact, prioritize feature development, and define a roadmap for agentic capabilities.
• Monitor open-source and industry trends in agent technologies to maintain alignment with emerging standards.

Qualifications and Experience

• Minimum 5 years in product management focusing on research products, agentic AI systems, or applied machine learning.
• In-depth knowledge of agent harness technologies including current commercial and open-source systems.
• Expertise in modern large language model agent architectures covering multi-agent systems, tool-augmented reasoning, memory/retrieval, orchestration, RAG, and complex workflow execution.
• Strong skills in designing agent evaluation metrics such as task success, recovery strategies, long-duration reliability, and distinguishing model vs. infrastructure limitations.
• Technical proficiency to actively shape architecture decisions including asynchronous execution, sandboxing, filesystem strategies, and platform-level tradeoffs.
• Ability to communicate fluently with both machine learning researchers and engineering teams.
• Demonstrated success shipping platform products that enhance reliability, performance, or functionality.

Preferred Experience

• Track record in developing and launching agent harness technologies.
• Active involvement with agent frameworks and orchestration tools in professional or personal projects.
• Exposure to enterprise-scale agent deployments, including multi-tenant management, tool permissioning, audit, and compliance.
• Understanding infrastructure challenges like on-premises operation, scalability, and security in restricted environments.
• Experience bridging research and product teams to translate emerging AI capabilities into delivered features.
• Background working near or within ML research or post-training teams.

Why Work at Cohere?

• Opportunity to influence AI adoption in major public institutions across Canada.
• Collaborate with leading researchers and engineers tackling complex AI challenges.
• Competitive salary, equity options, and support for professional growth.
• Flexibility with hybrid work environments and offices in multiple global cities.

Employee Benefits

• Weekly lunch stipend equivalent to $75 in local currency.
• Comprehensive health and dental coverage including mental health budgets.
• Retirement plans such as RRSP matching, 401K, or pension schemes.
• Parental leave with 100% top-up for up to 6 months available for either parent.
• Annual enrichment funds for arts, culture, wellness, workspace improvements, and learning.
• Generous vacation policy totaling 6 weeks (30 working days).
• Travel budget for remote employees to visit offices and annual company offsite events.

Work Environment

• Remote-friendly with offices in major cities worldwide.
• Office amenities include daily lunches, snacks, and social community events.
• Co-working support for remote employees to work alongside peers.
• Home office stipend of $500 for workspace setup.

Equal Opportunity and Recruitment Notes

Cohere embraces diversity and inclusivity, welcoming applicants from all backgrounds and committed to equitable hiring processes, including accommodations during recruitment. AI-enabled tools may assist candidate screening but do not limit recruiter discretion.

Compensation for this role ranges from $160,000 to $320,000 annually.