Product Manager - AI Platform

Supabase is a platform that supports over 7 million developers by providing integrated tools for database, authentication, storage, and APIs. They are seeking a Product Manager for their AI Platform to oversee the development and implementation of agent tooling, ensuring alignment across engineering, design, and leadership while addressing user needs and measuring product success.

Responsibilities:

• Rethink the PM job when most users are agents. Defaults, error messages, breaking changes, naming, deprecations, docs: every product decision behaves differently when an AI coding tool sees it before a human does. Apply that lens to your own roadmap and to product decisions across Supabase
• Talk to customers across the full spectrum. Indie developers shipping side projects, startups choosing an AI stack, enterprise teams running agents in production. Find the blockers and bring them back to the roadmap
• Own the problem statement and requirements behind every agent tooling bet. Define what we're building and why before engineering picks up a spec. Capture the evidence behind each decision, name the constraints, and give the team a target it can hit
• Decide what gets built, what gets deferred, and what gets cut. Every quarter you're choosing between agent plugin requests, MCP improvements, discoverability work, and reliability — and defending those calls when stakeholders push back
• Define how each launch is measured before it ships. Set the metric before engineering starts: developer activation, task completion, and AI eval scores. Track whether it moved after launch and use that to sharpen the next call
• Keep engineering, design, and leadership aligned. Agent tooling spans the MCP server, CLI, SDK, and dashboard. Communicate the roadmap, surface dependencies early, and keep stakeholders pointed in the same direction

Requirements:

• 7+ years of product management experience on developer tools, or an ex-founder with strong product instincts
• Technical enough to dive in: read an architecture doc, follow a design discussion, ask the right questions
• A clear point of view on what makes infrastructure agent-friendly
• Experience using AI coding tools and agent harnesses daily
• Opinions about the primitives developers need to build agent-native products
• Experience shipping something that put those opinions to the test
• Bias toward speed and willingness to ship something imperfect and learn
• Ability to know when to stop planning and start building
• Experience using the products on your roadmap: connecting to Supabase through the MCP server, building with the CLI, testing how agents respond to different tool configurations
• Experience using AI tools for research, drafting, and synthesising feedback
• Ability to work async by default