Role Overview
The Product Manager, Agent Harness & Modelling position at Cohere is a full-time opportunity based in Toronto with remote flexibility. This role suits an experienced professional ready to lead the execution layer that makes the North agent platform reliable and production-ready for enterprises. Operating at the intersection of agent loops, context engineering, and model-scaffolding co-evolution, the successful candidate will drive the architecture and development of advanced AI agent systems.
Key Responsibilities
Build and manage the roadmap for North's agent harness, encompassing agent loops, context management, tool orchestration, sandbox execution, and sub-agent delegation.
Act as the primary liaison between North engineering and Cohere's Modeling team to validate new harness capabilities prior to construction.
Oversee the agentic evaluation framework, ensuring compatibility across the North harness and Modeling training infrastructure.
Consult with enterprise clients to identify real-world agentic failures and translate those insights into actionable product and model requirements.
Collaborate with the agent expert modeling team to define target goals, prioritize essential model capabilities, and establish a clear roadmap for agentic modeling.
Monitor the open-source and commercial agent landscape to ensure the North architecture stays aligned with evolving industry standards.
Requirements
Minimum of 5 years of product management experience working with research PM roles, agentic AI systems, or applied machine learning products.
Comprehensive understanding of agentic harnesses, including current commercial and open-source solutions.
Familiarity with modern large language model agent architectures, such as multi-agent systems, tool-augmented reasoning, memory and retrieval, programmatic orchestration, RAG, and long-horizon execution.
Solid background in agentic evaluation design, including task completion measurement, failure recovery, and diagnostic abilities for model versus scaffolding gaps.
Technical depth to participate in implementation-level architecture decisions, review design documentation, and understand asynchronous execution patterns and sandboxed environments.
Ability to bridge machine learning research conversations and engineering architecture discussions fluently.
Demonstrated track record of shipping platform-layer products that improve reliability, performance, or capability.
Nice to Have
Direct experience developing and deploying leading agentic harnesses.
Active engagement as an experimental practitioner of agent frameworks in personal or professional projects.
Hands-on background with enterprise agentic deployments, including multi-tenant orchestration, permissioning, audit trails, and compliance.
Familiarity with infrastructure constraints in enterprise settings, such as on-premises environments, scalability challenges, and air-gapped operations.
Prior experience at the intersection of product and research, turning early model capabilities into final features.
Background working closely alongside a machine learning research or post-training team.
Skills and Tools
LLM
RAG
Agent Harnesses
Multi-Agent Systems
Context Engineering
Sandbox Code Execution
Agentic Evaluation Design
Salary and Benefits
Annual salary range of $160,000 to $320,000 USD.
Weekly lunch stipend of $75 or local currency equivalent.
Comprehensive health and dental benefits, alongside a dedicated mental health budget.
Retirement savings options including RRSP matching, 401K, and Pension Schemes.
Up to 6 months of 100% parental leave top-up for either parent.
Annual enrichment benefits covering arts, culture, fitness, wellness, and workspace improvements, plus education stipends for conferences and courses.
Generous paid time off providing 6 weeks of paid vacation (30 working days).
Travel budget for visiting other office locations and attending annual company offsites.
Home office setup stipend of $500 alongside co-working benefits for remote team members.
About the Company
Cohere is a leading security-first enterprise AI company that constructs foundational AI models and end-to-end products to address real-world business challenges. Headquartered in Toronto with major international offices, the global team consists of researchers, engineers, and designers dedicated to advancing AI adoption. Cohere also develops North, an enterprise agentic AI platform that enables organizations to securely deploy agents and automation within their own infrastructure.
How to Apply
Interested candidates should visit the official Cohere careers page to submit an application for this position.
