Security Engineer, Trust & Safety - Applied AI

Meta•Washington, DC

2d•$154,000 - $217,000

About The Position

Meta is seeking a Security Engineer with Trust & Safety expertise to join its Applied Artificial Intelligence (AAI) organization. The role applies deep expertise in online abuse and adversarial behavior—including child safety, human exploitation, fraud and scams, violent extremism, and coordinated inauthentic behavior—to advance AI capabilities in Trust & Safety. The engineer uses AI tools and agents to accelerate work and applies domain expertise to improve models where they fall short. The position focuses on abuse driven by behavior, context, and scale rather than software defects. It includes developing robust evaluations for internal use cases, architecting Trust & Safety capabilities and evaluation infrastructure, and shaping methodologies that make AI model improvement systematic and scalable. The role partners across teams and organizations, leads technically complex work, mentors engineers, and shares findings through internal publications and broader community contributions.

Requirements

B.S. or M.S. in Computer Science, Cybersecurity, or a related field, or equivalent experience; alternatively, a Ph.D. plus 2 years of hands-on engineering experience.
5+ years of hands-on engineering experience in Trust & Safety, integrity, fraud and abuse, or security investigations.
Demonstrated depth in at least one abuse domain: child safety and youth protection; dangerous organizations and violent extremism; fraud and scams; coordinated inauthentic behavior and information operations; or content policy enforcement.
Experience building detection, enforcement, or evaluation systems in an abuse domain, in addition to operating or investigating within one.
Experience analyzing and interpreting large-scale data, including relational databases and SQL, to identify adversarial behavior patterns and quantify trends.
Experience investigating and acting on high-impact threats and online threat actors.
Coding experience in at least one language such as PHP, Python, or Java, or a similar language.

Nice To Haves

Ongoing AI skill development, such as prompt or context engineering and agent orchestration, and staying current with emerging AI technologies.
Experience with campaign-level investigation and attribution of coordinated adversary networks.
Experience implementing responsible and ethical AI practices, such as risk assessment, bias mitigation, and quality and accuracy reviews.
Experience adversarially testing conversational or agentic systems, including jailbreaks, social engineering, prompt injection, and policy-exception pressure.
Experience authoring or interpreting enforcement guidelines at scale, including coded terminology and evolving slang.
Experience with age assurance and age-signal inference, including calibrated confidence and mode switching.
Contributions to the Trust & Safety or security community, such as original research, tools, conference presentations, or publications.
Experience developing and evaluating classifiers for abuse, fraud, or policy enforcement.
Demonstrated ability to integrate AI tools to optimize or redesign workflows and drive measurable impact, such as efficiency gains or quality improvements.
Experience improving AI model performance through expert feedback, red-teaming, or evaluation design.

Responsibilities

Identify and solve ambiguous problems within a Trust & Safety domain, driving solutions that affect multiple teams.
Architect Trust & Safety capabilities and evaluation infrastructure, establish standards across teams, and drive platform-level decisions for AI-driven solutions at Meta's scale.
Lead strategy and methodology for abuse problems AI cannot yet reason about, including age-aware policy boundaries, adaptive adversaries, campaign-scale abuse, and policy interpretation under pressure.
Define methodologies that translate Trust & Safety expertise into systematic and scalable AI model improvement.
Drive cross-organizational partnerships and influence technical direction at the intersection of Trust & Safety and AI.
Break problems into scoped workstreams, own the most technically complex pieces, and enable others to execute the overall solution.
Mentor engineers across the organization, build community, and model Trust & Safety engineering practices and Meta values.
Disseminate findings through internal publications, knowledge sharing, and contributions to the broader community.