Product Manager - Agent Harness & Modelling

About Cohere

Cohere is a leading enterprise AI company prioritizing security, developing advanced foundational AI models and comprehensive products designed to tackle practical business challenges. Our global team of researchers, engineers, and designers is dedicated to enhancing our models' capabilities and delivering exceptional value to customers. Our headquarters are in Toronto, with offices worldwide including London, New York, San Francisco, Montreal, Paris, Berlin, and Seoul.

About North

North is Cohere's groundbreaking agentic AI platform engineered to deploy AI agents and automations securely within organizational infrastructures. It boosts employee productivity by streamlining workflows, automating repeat tasks, and revealing actionable intelligence while maintaining data privacy and compliance. North integrates advanced generative and search models with customizable tools to foster large-scale innovation.

Role Overview

We seek a Product Manager to lead our Agent Harness execution layer within the North platform. This role involves ownership of the core runtime systems enabling agent reliability, including tool orchestration, parallel workflows, sub-agent delegation, sandbox execution, and failure recovery. The position requires deep technical collaboration with engineering and synergy with the Modeling team to continuously evolve both the harness and the AI models.

Responsibilities

• Develop and manage the roadmap for North's agent harness features such as agent loops, context engineering, tool orchestration, sandboxed executions, and delegation mechanisms.
• Act as the main liaison between North engineering and Cohere's Modeling teams, ensuring feature validation before development and collaborative problem-solving.
• Oversee North's agent evaluation framework to align product functionality with research training pipelines, bridging technical and research domains.
• Engage with enterprise customers to identify real-world challenges in agentic AI and translate insights into product and model enhancements.
• Collaborate closely with modeling experts to establish impactful model goals, set priorities for model capabilities and evaluations supporting product objectives.
• Maintain awareness of developments in open-source and commercial agent ecosystems to inform architectural alignment and adoption strategies.

Requirements

• A minimum of 5 years' experience in product management focused on research-oriented AI, agentic AI platforms, or machine learning-driven products.
• Profound understanding of agent harness architectures, including current leading proprietary and open-source platforms.
• Strong knowledge of modern large language model (LLM) agent ecosystems such as multi-agent coordination, tool augmentation, memory/retrieval systems, programmatic orchestration, retrieval-augmented generation (RAG), and handling long execution horizons.
• Expertise in designing evaluation methods for agent systems, including metrics for task success, failure handling, long-term reliability, and diagnostic differentiation between model and scaffolding issues.
• Technical depth sufficient to partake in architectural decisions, including asynchronous execution, sandboxed environments, filesystem design, and production platform trade-offs.
• Ability to navigate fluently between machine learning research dialogues and engineering architecture discussions.
• Demonstrated success releasing platform-level products that improve reliability, performance, or capabilities.

Preferred Qualifications

• Hands-on experience building and delivering leading agentic AI harnesses.
• An active user and contributor to agent frameworks with up-to-date knowledge of open source harnesses and orchestration tools.
• Experience deploying enterprise-grade agents considering multi-tenant orchestration, permissioning, audit compliance, and regulatory factors.
• Familiarity with on-premise infrastructure restrictions, scalability challenges, and operational compromises for complex agent workloads in controlled environments.
• Proven track record translating research model advances into tangible product features.
• Background interacting with or working within dedicated machine learning research or post-training teams.

Why Work With Cohere?

• Make a significant impact on leading public institutions adopting advanced AI.
• Collaborate with top researchers and engineers tackling hard ML problems.
• Receive competitive pay, equity participation, and access to career growth opportunities.
• Enjoy a flexible hybrid work culture with global office presence.

Employee Benefits

• Weekly lunch stipend of $75 (or local equivalent).
• Comprehensive health and dental insurance with mental health funding.
• Retirement plans including RRSP matching, 401(k), and pension schemes.
• Fully paid parental leave up to 6 months for either parent.
• An annual enrichment budget covering arts, fitness, wellness, work environment, educational courses, conferences, and coaching.
• Generous vacation totaling 30 working days per year.
• Travel budget for remote employees to visit other offices plus an annual company offsite event.

Work Environment

• Flexible remote-friendly culture with offices in multiple global cities.
• Office amenities include daily lunch service, snacks, and community events.
• Co-working stipends for remote employees to work alongside peers locally.
• $500 home office setup reimbursement per employee.

Additional Information

We encourage all qualified candidates to apply regardless of exact match to listed experiences. Our hiring process supports accommodations upon request to ensure an inclusive and equitable recruitment experience. We utilize AI tools to aid screening but ensure human review of candidates. Beware of recruitment scams; official communication is only via company domains.

Compensation

Salary range: $160,000 to $320,000 annually.