Software Engineer, Guardian
Core
Design, build, and operate real-time production systems (Auto-review and Guardian) to evaluate AI agent actions, detect harmful behavior, and ensure safe recovery or escalation.
Role type
Senior IC software engineer (AI safety & alignment)
Builds
Real-time production systems for AI agent safety, oversight, and safe recovery
Domain
AI safety, alignment research, and large language model (LLM) systems
Deliverable
production ML models
Required skills
Backend services design, API development, infrastructure design, AI agent/LLM integration, adversarial testing, system reliability, product-engineering judgment, cross-functional collaboration
Preferred skills
Experience translating research prototypes into scalable products, comfort with ambiguity, initiative
Technologies
None explicitly stated
Responsibilities
Design and build Auto-review and Guardian platform components; develop evaluations and adversarial tests for safety performance; integrate Guardian capabilities into agent experiences; investigate failures to improve reliability; partner with alignment researchers to translate prototypes into scalable capabilities
Seniority
Senior, hands-on IC
