- Redmond, Washington, United States +1 more
- Principal
- $143K–$275K per year
- 3+ years
Design and execute white-box attacks against advanced AI models, checkpoints, post-training systems, reward functions, data pipelines and evaluation environments.
Locations (2)
- Redmond, Washington, United States
- United States
Compensation
Research Sciences IC5 - The typical base pay range for this role across the U.S. is USD $142,800 - $274,800 per year. There is a different range applicable to specific work locations, within the San Francisco Bay area and New York City metropolitan area, and the base pay range for this role in those locations is USD $188,000 - $304,200 per year.
Responsibilities
- Design and execute white-box attacks against advanced AI models, checkpoints, post-training systems, reward functions, data pipelines and evaluation environments.
- Investigate adversarial behavior including jailbreaks, harmful permissiveness, reward hacking, deceptive behavior, capability elicitation, privacy leakage, data poisoning and model supply-chain compromise.
- Build high-fidelity experiments and automated attack suites that reproduce realistic threats and run continuously against new checkpoints and system configurations.