Researcher – Safety Training, National Security

Jobtailor; California (MO); On-site; USD 140,000 - 210,000; Full time

Job summary

Jobtailor is seeking an AI safety researcher to advance methods for safety training, RLHF, and model robustness. You will develop evaluations, identify failure modes, and collaborate with cross-functional teams to support safe deployment for national security applications.

The role requires 4+ years in AI safety research, a CS/ML degree, and active TS/SCI clearance. You will train and evaluate models used in government contexts and contribute to policy-aligned outcomes.

Qualifications

• 4+ years of AI safety research experience, including RLHF, adversarial training, or robustness.
• A degree in computer science, machine learning, or a related field
• Strong deep learning research or engineering skills
• Experience improving model safety for deployment
• Active TS/SCI clearance or equivalent
• Motivation to support the responsible use of AI in safety-critical settings

Responsibilities

• Research and implement methods for safety training, reinforcement learning, and adversarial robustness
• Develop evaluations and identify model failure modes
• Use evaluation findings to improve training
• Work with research, engineering, security, and policy partners to support safe, reliable deployment
• Train and evaluate models for U.S. government use, with a focus on national security applications
• Advance safety post-training and robustness while helping models follow nuanced policies and preserve usefulness and capabilities

Skills

AI safety research

RLHF

Adversarial training

Deep learning research

TS/SCI clearance

Education

CS/ML degree or related field

Tools

Reinforcement Learning

Adversarial Robustness

Model Evaluation

Model Safety Improvement

Deep Learning Engineering

Job description

• Research and implement methods for safety training, reinforcement learning, and adversarial robustness


• Develop evaluations and identify model failure modes


• Use evaluation findings to improve training


• Work with research, engineering, security, and policy partners to support safe, reliable deployment


• Train and evaluate models for U.S. government use, with a focus on national security applications


• Advance safety post-training and robustness while helping models follow nuanced policies and preserve usefulness and capabilities

Requirements

• 4+ years of relevant AI safety research experience, including RLHF, adversarial training, or robustness


• A degree in computer science, machine learning, or a related field


• Strong deep learning research or engineering skills


• Experience improving model safety for deployment


• Active TS/SCI clearance or equivalent


• Motivation to support the responsible use of AI in safety-critical settings

Core Competencies

Demonstrates expertise in AI safety research, focusing on reinforcement learning, adversarial training, and model robustness. Proven ability to collaborate with cross-functional teams to enhance model safety for national security applications.

Highest-signal resume keywords

• AI Safety Research Experience


• Reinforcement Learning from Human Feedback (RLHF)


• Adversarial Training


• Deep Learning Research


• Active TS/SCI Clearance

Hard Skills

• Reinforcement Learning


• Adversarial Robustness


• Model Evaluation


• Model Safety Improvement


• Deep Learning Engineering

Soft Skills

• Collaboration


• Motivation for Responsible AI Use

Certifications & Qualifications

• Active TS/SCI Clearance

Industry Keywords

• National Security Applications


• Safety-Critical Settings


• Model Failure Modes


• Policy Compliance