10a Labs logo

Model Safety Investigations Specialist

10a Labs

Washington, DC; San Francisco, CA; New York City, NY; Denver, CORemotePosted 2 days ago$120,000 - $150,000/year
ContractMid LevelTech & Data (Health IT)Others
Original posting ↗

Summary

This is a 9-month position, with the potential for extension. We are seeking Model Safety Investigations Specialists to support inbound model safety investigations. Specialists will investigate reports of unexpected model behavior, assess the available technical evidence, and document what it shows. Specialists will follow established procedures and escalate urgent or ambiguous cases to the designated investigation lead.

Job description

About the Role

This is a 9-month position, with the potential for extension. We are seeking Model Safety Investigations Specialists to support inbound model safety investigations. Specialists will investigate reports of unexpected model behavior, assess the available technical evidence, and document what it shows. Specialists will follow established procedures and escalate urgent or ambiguous cases to the designated investigation lead.

In this role, Specialists will track case volume, complexity, and handling effort, and flag where guidance or tooling needs improvement.

This role may involve exposure to sensitive or distressing content. Strong discretion, sound judgment, and careful handling of user information are essential.

In This Role, You Will

  • Triage inbound reports, check for duplicates and relevant history, and gather the minimum evidence needed using approved tools.
  • Use available logs and case evidence to investigate unexpected model behavior and document what happened and what remains unclear.
  • Review cases in context and apply severity and labeling guidance consistently. Distinguish observed behavior and reported harm from hypotheses about cause, intent, or prevalence.
  • Write case summaries that capture evidence, uncertainty, rationale, and a recommended next step. Escalate time-sensitive or high-severity cases promptly.
  • Label cases, participate in calibration and quality reviews, and flag recurring patterns, ambiguous guidance, or detection gaps.
  • Track case complexity, handling time, and escalations so the team can assess coverage and refine longer-term staffing needs.

Required Qualifications

  • 2 - 5 years of hands-on experience investigating sensitive technical cases in trust and safety, incident response, fraud, abuse, or a similar operational setting.
  • Ability to reconstruct scenarios from incomplete or conflicting evidence and write a concise, defensible summary another reviewer can verify.
  • Ability to apply policy or severity guidance consistently, recognize when a case exceeds your authority, and escalate without overstating the evidence.
  • Can manage a queue without sacrificing accuracy, incorporate calibration feedback, and follow strict privacy and access controls.
  • Comfortability interpreting technical information, such as logs, and piecing together evidence from multiple sources.

Preferred Qualifications

  • Experience reviewing AI conversations or investigating model behavior, such as jailbreaks, unsafe outputs, or unexpected agent actions.
  • Experience with case labeling, reviewer calibration, or quality assurance. Familiarity with basic SQL or approved AI-assisted analysis tools is helpful.

Requirements

  • 2 - 5 years of hands-on experience investigating sensitive technical cases in trust and safety, incident response, fraud, abuse, or a similar operational setting
  • Ability to reconstruct scenarios from incomplete or conflicting evidence and write a concise, defensible summary another reviewer can verify
  • Ability to apply policy or severity guidance consistently, recognize when a case exceeds your authority, and escalate without overstating the evidence
  • Can manage a queue without sacrificing accuracy, incorporate calibration feedback, and follow strict privacy and access controls
  • Comfortability interpreting technical information, such as logs, and piecing together evidence from multiple sources
  • Experience reviewing AI conversations or investigating model behavior, such as jailbreaks, unsafe outputs, or unexpected agent actions
  • Experience with case labeling, reviewer calibration, or quality assurance
  • Familiarity with basic SQL or approved AI-assisted analysis tools

Benefits

  • Health, Dental & Vision
  • Generous PTO
  • Paid Holidays
  • 401(k) Plan
  • Fully Remote

Company overview

10a Labs is a specialized AI security company that serves as the safety and threat-intelligence layer for frontier AI labs, AI unicorns, Fortune 10 companies, and leading global technology platforms. Founded in 2021 and headquartered in San Francisco, the company provides critical security services focused on protecting advanced AI systems from emerging threats. Their core offerings include adversarial red teaming, independent model evaluations, and comprehensive threat intelligence collection. Through proprietary technologies like Spartan for scaled red teaming and Scout for classification systems, 10a Labs enables engineering, safety, and security teams to identify vulnerabilities, assess risks, and deploy AI systems safely. The company specializes in detecting and mitigating threats across various domains including CBRNE (chemical, biological, radiological, nuclear, and explosive), cyber misuse, harmful manipulation, model autonomy, and violent activities. With a team of experts in AI security, data science, and threat investigation, 10a Labs plays a crucial role in the evolving landscape of AI safety, helping organizations comply with regulatory frameworks like the EU AI Act while maintaining robust security postures against sophisticated adversarial attacks.

Explore more jobs of your matchFind more biopharma roles with GetBioJob.