Research Engineer, Frontier Safety Mitigations
$174k - $252k • Mountain View, CA, USA • Full-time
Posted 1mo ago
Job Location
Mountain View, CA, USA
Tech Stack
Remote Work Policy
On-site
Employment Type
Full-time
Categories
AI Research Engineer
About the job
Our team focuses on de-risking model launches by defending against critical misuse domains such as Cybersecurity, CBRNE, and Harmful Manipulation. This involves building novel evaluations, red-teaming, researching and deploying advanced mitigations (both in-model and out-of-model), and monitoring emerging risks to ensure robustness while enabling beneficial use. We are looking for a research engineer to help build the next generation of safety mitigations for frontier models, focusing on robust, end-to-end defenses against severe risks. This work feeds directly into DeepMind's Frontier Safety Framework commitments, requiring comfort with approaching new research questions and implementing technical solutions at scale.
Responsibilities
- Build advanced classifiers and data pipelines to detect misuse, from automated evaluation to rapid model iteration.
- Develop cross-context monitoring systems to detect coordinated harms using novel signal aggregation methods.
- Implement semi-automated account-level response systems to detect, track, and apply strikes against malicious actors.
- Evaluate and secure agentic AI systems by developing threat models, creating testing environments, and deploying mitigations against frontier-level agentic hacking.
- Advance research in automated red-teaming and adversarial robustness using multi-turn/agentic attacks to uncover misuse vulnerabilities.
Requirements
- Bachelor's degree or equivalent practical experience.
- 5 years of experience with software development in one or more programming languages.
- 3 years of experience testing, maintaining, or launching software products.
- 1 year of experience with software design and architecture.
- Master's degree or PhD in Computer Science or a related technical field (preferred).
- Experience in Large Language Model (LLM) development, fine-tuning, or safety evaluation methodologies (preferred).
- Experience taking research from concept to product (preferred).
- Experience in Safety and Alignment (preferred).
Benefits
- 15% bonus target
- Equity
- Benefits
