Responsibilities

E-commerce's Governance and Experience (GNE) is a global team responsible for maintaining a safe and trustworthy marketplace for users, sellers, and creators. The Seller & Creator Service Strategy & Automation team owns the service strategy and the automation capability behind it — deciding what sellers and creators should be able to resolve without a human, and building the agent capability that delivers it across every market. We are looking for an automation program manager to own that resolution capability — moving it from prompt-and-knowledge configuration to executable, versioned Skill contracts backed by MCP tools, and proving out the gains market by market through controlled experiments. Roles & Responsibilities - Define what the agent can resolve. Turn seller and creator service scenarios into executable Skill contracts — what's in scope, what the agent may act on, and where it must stop. - Shape the tool layer. Work with product and engineering to define the capabilities the agent needs to actually complete a task, not just answer a question. - Set the automation boundary. Partner with policy and regional service owners on which scenarios are automated, which stay human-assisted, and where the handoff sits. - Launch and prove it. Take solutions live through controlled experiments, own the readout, and make the call on whether a capability is ready to scale. - Scale across markets. Extend and adapt capabilities across regions, business models, and both seller and creator surfaces. - Own the automation metrics. Break the broad automation rate down to its component metrics, set targets on each, and keep measurement definitions aligned ahead of rollout. - Keep it improving. Monitor resolution quality in production, attribute failures to their root cause, and turn findings into the next round of changes.

Qualifications

Minimum Qualification(s) - Bachelor's degree or higher. - 3+ years in service automation, conversational AI, or LLM application delivery — with at least one agent or chatbot capability you owned end to end from design to launch. - Hands-on understanding of LLM agent mechanics: tool calling, ReAct-style loops, context construction, and where retrieval should and should not substitute for a tool call. Prompt engineering is table stakes, not the job. - Ability to read and challenge experiment results: A/B design, significance, and the difference between in-scenario and overall denominators. SQL or equivalent self-serve data capability. - Comfortable specifying interfaces with engineering — API/tool semantics, batch behavior, failure and idempotency handling — without owning the implementation. - Strong problem-solving and change-management skills; able to drive results across product, algorithm, policy, and regional service teams in a deadline-driven, multi-market environment. Preferred Qualification(s) - Experience operating across multiple regional markets. - eCommerce marketplace experience, or experience serving creator/content ecosystems, is a plus.

Job Information