Staff Software Engineer, Ads Agent Infra

Role overview Set technical direction and build infrastructure for stateful AI agents used across advertising and monetization products. The focus is on reliable distributed execution, shared agent interfaces, and operational controls that help engineering teams deploy resilient workflows at scale.

Responsibilities - Architect and maintain durable execution platforms that can recover stateful agent workflows after failures. - Create standardized protocol servers, SDKs, and runtime interfaces for teams building agents. - Establish observability for language-model systems, including tracing, prompt versioning, and token-cost controls. - Write foundational design documents and align infrastructure, security, and product stakeholders around technical strategy. - Automate testing, evaluation, and grading pipelines for agent skills and memory services. - Use AI to speed up design, documentation, and coding while verifying correctness and quality.

Requirements - At least eight years of software engineering experience, including distributed systems, microservices, and durable execution technologies. - Hands-on knowledge of LLM agent frameworks and runtime internals. - Experience setting direction for platform abstractions and aligning teams across infrastructure, security, and product. - Background in service identity, multi-tenant security, and observability or cost governance for LLM systems. - Demonstrated use of AI to improve day-to-day engineering work, with careful testing, source checking, and peer review of AI-assisted outputs. - Strong ownership and judgment when handling sensitive data and making technical decisions. - Bachelor’s or master’s degree in computer science or a related field, or equivalent experience.

Benefits and work setup The position requires in-person collaboration once or twice each quarter and must be within commuting distance of offices in San Francisco or Palo Alto. Relocation assistance is not available.