Adversarial AI Safety Engineer (Open Models)

Reflection AI Ltd; New York (NY); On-site; USD 150,000 - 210,000; Full time

Benefits offered by this job

Top-tier pay

Stock options

Health insurance

Meals provided

Parental leave

Unlimited PTO

Team events

Job summary

Reflection AI Ltd is seeking a leader to own the red-teaming and adversarial evaluation for its models, ensuring safety under deployment. You will work closely with the Alignment team to translate findings into practical guardrails and verify that every release meets risk thresholds before shipping.

As a senior contributor, you will develop scalable safety benchmarks and stay ahead of vulnerabilities with dynamic adversarial testing methodologies.

Qualifications

• Graduate degree or equivalent practical experience in AI Safety.
• Deep technical understanding of LLM safety, adversarial attacks, red-teaming, and interpretability.
• Strong software engineering skills with automated evaluation pipelines or large‑scale ML systems.

Responsibilities

• Own the red-teaming and adversarial evaluation pipeline for Reflection’s models.
• Translate safety findings into guardrails with the Alignment team.
• Validate releases meet risk thresholds before shipping.
• Develop scalable, automated safety benchmarks.
• Research jailbreaking techniques and defenses to stay ahead of vulnerabilities.

Skills

LLM safety

Adversarial testing

Red-teaming

Evaluation pipelines

RLHF/RLAIF

Education

MS/PhD in CS/ML

Tools

Python

PyTorch

RL frameworks

Job description

Reflection AI Ltd is seeking a leader to own the red-teaming and adversarial evaluation for its models, ensuring safety under deployment. You will work closely with the Alignment team to translate findings into practical guardrails and verify that every release meets risk thresholds before shipping.

As a senior contributor, you will develop scalable safety benchmarks and stay ahead of vulnerabilities with dynamic adversarial testing methodologies.