Remote job
AI Safety Engineer
Job details
About this role
Role overview
An AI Safety Engineer focused on identifying, measuring, and mitigating risks associated with advanced machine learning systems. The position is full-time and can be performed remotely or from a San Francisco office, with work centered on building safeguards that keep AI behavior aligned with intended goals across the model lifecycle.
Responsibilities
- Design and run evaluations that surface unsafe, biased, or unintended model behaviors - Build monitoring, red-teaming, and intervention pipelines for production AI systems - Collaborate with research and engineering teams to translate safety findings into model and system improvements - Document safety incidents, risk assessments, and mitigation decisions for cross-functional review - Contribute to internal policies, standards, and tooling that govern responsible AI development
Requirements
- Hands-on experience with machine learning frameworks and model evaluation workflows - Familiarity with AI risk categories such as misuse, jailbreaks, hallucinations, and bias - Strong programming skills in languages commonly used for ML and data analysis - Ability to write clear, structured technical documentation - Comfort working in a fast-moving environment with cross-disciplinary teams
Nice to have
- Background in adversarial machine learning, interpretability, alignment research, or closely related areas - Experience contributing to safety benchmarks, evaluation suites, or published research
Benefits and work setup
- Full-time engagement with flexibility to work remotely or from a San Francisco office