AI Engineer, Ecosystem

Responsibilities

• Own Brex's MCP Apps for OpenAI, Claude, Slack, and other embedders, including roadmap decisions across core and long-tail partners.
• Define the technical, UX, and reliability standards for how Brex surfaces and operates within external AI assistant ecosystems.
• Design and maintain secure, auditable integration patterns, including permissions, read/write access, and acting on behalf of users across third-party environments.
• Build and operate production-grade integrations that remain resilient amid evolving platforms, unpredictable clients, and shifting constraints.
• Partner cross-functionally with Security, Platform, Sales, Enablement, and internal agent teams to ensure integrations are safe, valuable, and well-supported.
• Serve as the center of excellence for reusable AI capabilities across Brex, defining and evangelizing standards, patterns, and shared components that power AI functionality both within the Brex platform and across external ecosystem integrations.

Requirements

• Experience building and owning production software, including systems that integrate with or depend on third-party platforms.
• Demonstrated technical ownership for user-facing products with real external customers.
• Deep experience with identity, authentication, and secure data access across system boundaries, including OAuth and role-based access control (RBAC) for permissions.
• Proven ability to ship and operate reliable integrations with external APIs and clients outside your control.
• Strong product judgment, written communication, and end-to-end ownership, including accountability for quality and launch readiness.

Bonus Points

• Experience publishing and maintaining apps or integrations in third-party marketplaces (iOS, Slack, Shopify, GitHub, Chrome, etc.).
• Prior work on administrative or platform-side tooling for ecosystems.
• Familiarity with conversational or agentic interfaces in production environments.
• Experience building evaluations to measure and improve the quality, reliability, and safety of AI-powered features.