Senior Member of Technical Staff - Model Safety

Xcede; San Francisco (CA); On-site; USD 180,000 - 240,000; Full time

Job summary

Xcede partners with a frontier AI research company to recruit a Member of Technical Staff focused on AI Safety. You will lead red-teaming, design scalable safety evaluation pipelines, and work at the intersection of model evaluation, alignment, and software engineering.

You will collaborate with alignment researchers, set safety thresholds for releases, and develop state-of-the-art jailbreaking defenses to stay ahead of threats.

Qualifications

• Deep expertise in LLM safety, red teaming, adversarial attacks, interpretability, or model evaluation.
• Strong software engineering skills and experience building large-scale ML systems or automated evaluation infrastructure.
• Experience with AI alignment techniques, RLHF, RLAIF, or related safety methodologies is highly desirable.
• MS, PhD, or equivalent practical experience in Machine Learning, AI Safety, Computer Science, or a related field.

Responsibilities

• Lead red-teaming and adversarial evaluation efforts to uncover model vulnerabilities, misuse risks, and alignment gaps.
• Design and build scalable safety evaluation frameworks and automated testing pipelines.
• Partner closely with alignment researchers to translate safety findings into production guardrails.
• Act as a key decision-maker in determining whether model releases meet safety thresholds.
• Research and implement state-of-the-art jailbreaking techniques and defenses to stay ahead of emerging threats.
• Develop dynamic safety benchmarks that evolve alongside increasingly capable AI systems.

Skills

LLM safety

Red teaming

Adversarial attacks

Model evaluation

Large-scale ML systems

AI alignment methods (RLHF/RLAIF)

Education

MS/PhD in ML/CS

Tools

Python

ML infra

CI/CD

Job description

We're partnering with a frontier AI research company on a search for a Member of Technical Staff focused on AI Safety.

The company is building next-generation open-weight foundation models with a mission to make advanced AI broadly accessible. Their team includes researchers, engineers, and operators from some of the world's leading AI labs and technology companies, working on the frontier of model capabilities, alignment, and deployment.

This is an opportunity to help define how advanced AI systems are evaluated, stress-tested, and safely deployed. You'll be working at the intersection of adversarial research, red teaming, model evaluation, alignment, and software engineering—helping ensure next-generation foundation models are robust, reliable, and secure before they reach the world.

Overview:

• Lead red-teaming and adversarial evaluation efforts to uncover model vulnerabilities, misuse risks, and alignment gaps
• Design and build scalable safety evaluation frameworks and automated testing pipelines
• Partner closely with alignment researchers to translate safety findings into production guardrails
• Act as a key decision-maker in determining whether model releases meet safety thresholds
• Research and implement state-of-the‑art jailbreaking techniques and defenses to stay ahead of emerging threats
• Develop dynamic safety benchmarks that evolve alongside increasingly capable AI systems

Skills required:

• Deep expertise in LLM safety, red teaming, adversarial attacks, interpretability, or model evaluation
• Strong software engineering skills and experience building large-scale ML systems or automated evaluation infrastructure
• Experience with AI alignment techniques, RLHF, RLAIF, or related safety methodologies is highly desirable
• MS, PhD, or equivalent practical experience in Machine Learning, AI Safety, Computer Science, or a related field
• Ability to operate in a high-agency, fast-moving environment and make high-stakes technical decisions