Owns the interfaces and abstractions, MCP servers, agent skills, and dashboard assistants that developers and AI agents use to interact with a platform. Drives an eval-first approach to AI tooling quality.
Differs from context-engineer (which curates retrieval and prompt context for models): this role builds the actual tools and MCP servers agents call. Differs from ai-conversation-designer (which designs conversational flows): this is developer-facing tooling infrastructure.
Supabase's AI Tooling Engineer builds MCP servers, agent skills, and the dashboard assistant with an eval-first approach. No public salary range disclosed; pay omitted rather than estimated.
