Product Manager, Agent Platform

OpenRouter is an AI routing and infrastructure company that helps developers and enterprises access, manage, and optimize AI models across providers. The company is seeking its first Product Manager for the Ori Agent Platform to define its strategy, roadmap, positioning, and growth from private testing to a business. The role involves launching developer tools, improving multi-model agent experiences, and coordinating with design, go-to-market, and platform teams.

Responsibilities:

• Own the agent platform roadmap end to end across Harness, Code, Eval, Intern, and Desktop, from the CLI today to teams of interns and multi-agent deployments
• Set the strategy. Define the target user and competitive positioning for each surface, and decide where Ori should invest across coding, cloud, and desktop agents. Position Ori as an open reference implementation to inspire the multi-model market
• Take new tools from waitlist to general availability. Run private testing and design partners, drive internal dogfooding, and instrument adoption, retention, and spend
• Make model fungibility real. Model switching, provider fallbacks, guaranteed tool dependencies, org guardrails, and eval-driven model selection that work in any harness and on any surface
• Own agent experience, from install and auth to CLI UX, harness configuration, secrets, connections, and the app interfaces. You’ll help us set the quality bar for new consumer-grade experiences
• Close the loop with go-to-market and the platform. Turn waitlist, tester, and enterprise signal into the roadmap, and push the API and SDK gaps Ori exposes into the core product

Requirements:

• 5+ years of product management or founder experience building developer tools, agent products, or infrastructure for technical users
• A zero-to-one track record. You have taken a product from prototype to paying users and are comfortable with the ambiguity and hard tradeoffs that come with it
• Daily fluency with coding agents. You use tools like Claude Code, Codex, Hermes, or OpenCode and have specific opinions about what makes an agent loop, a harness, or a tool integration good
• Strong technical judgment. You read code, run the CLI, inspect traces, and reason with engineers about agent loops, tool calling, context management, evals, and the cost, latency, and quality tradeoffs between models
• Strong product taste for developer tools. You've shipped technical products where interaction details matter, and can carry that judgment into new interfaces
• High agency and a bias toward action. You talk to users, prototype, ship, and iterate on a small team with fluid boundaries. You write clearly and often
• Founder or early employee of a developer tools or agent company
• Experience with multi-agent systems, agent skills or memory, or evals and benchmarking products
• Experience shipping Slack apps or native desktop apps, or growing a product with both developer and non-developer users
• Existing user of OpenRouter or active side projects in AI tooling