Minimum qualifications:- Bachelor's degree in Computer Science, Machine Learning, or a related technical field, or equivalent practical experience.
- 5 years of experience in engineering and agentic assistance, including software development in Python.
- Experience working in a frontier AI research and development environment.
- Experience working in a professional software engineering or research team environment.
- Experience working with technical stakeholders.
- Experience in frontier model risk.
Preferred qualifications:- Experience of engineering or product design for AI tools or assistants, especially those focused on ML Research and Development (R&D).
- Experience with cybersecurity detection and response.
- Experience with collaborating or leading an applied ML project.
- Experience with Large Language Model (LLM) training and inference.
- Knowledge of AI control, chain-of-thought and other monitoring, faithfulness and monitorability and related research areas.
About the jobOur team develops monitoring and control for potentially misaligned AI to mitigate risks of extreme harms. Currently, this primarily involves: designing, building, and testing monitors for potentially dangerous behaviours; developing and implementing response policies to preserve AI usefulness while mitigating risks; and foreseeing ways in which our control tools might be bypassed or degraded. We are looking for an engineer who can rapidly iterate to solve never-before-seen problems with creativity and thoroughness.
The Loss of Control team contributes to a defense in depth against the risk of misaligned AI systems being deployed. We take the possibility of very advanced AI seriously. We don't think control is a suitable alternative to alignment in the limit of advancing intelligence. But while AI remains effectively monitorable, we think that control is an important part of an overall strategy for building safe AI.
We are looking for a research engineer for the Frontier Safety Loss of Control team within the AGI Safety and Alignment Team based in either San Francisco or London.
In this role, the core responsibility is to help Google prepare for the internal use of potentially misaligned AI systems. That means building defense-in-depth against AI that might persistently pursue goals that users and system developers did not intend.
US: $174000 - $253000 (USD) 15% bonus target bonus equity benefits
Learn more about benefits at Google .
Responsibilities- Identify potential harms from misaligned agents and develop strategies for detection and prevention.
- Implement technical controls to monitor agent thoughts, behaviour, and respond to mitigate potential harms.
- Integrate various agent behaviour signals from across the organisation to inform response policies.
- Conduct adversarial testing of controls.
- Work with internal product teams to ensure that control systems are adopted over all high-risk AI surfaces.
