Privacy Security Engineer, Privacy & Data Protection - Applied AI

The problems that matter here are the ones a capable engineer cannot solve by reading the code in front of them: which jurisdiction's deadline applies to a deletion request, whether a derived dataset clears an anonymity threshold, where a deletion has to cascade. Models already handle the privacy obligations that are stated as ordinary good engineering practice, so the work is the obligation they cannot derive.

This role is for privacy engineers who want to operate at the intersection of deep privacy expertise and frontier AI, not just applying AI to privacy work but shaping what it is capable of.

Responsibilities

Identify and solve ambiguous problems within the privacy and data protection space where the solution and requirements are initially unknown, driving solutions that affect multiple teams Architect privacy capabilities and evaluation infrastructure that sets the standard across teams; drive platform-level decisions for AI-driven privacy solutions at Meta's scale Lead the strategy and methodology for the privacy problems AI cannot yet reason about: jurisdictional and regulatory obligations, de-identification standards, data lifecycle controls, and data access reasoning Shape how privacy expertise translates into AI capabilities: define the methodologies that make model improvement systematic and scalable Drive cross-organizational partnerships that define how privacy and AI intersect: influence technical direction beyond the immediate team Break down problems into smaller scoped work streams, owning the most technically complex pieces while enabling others to execute against the overall solution Mentor other engineers across the organization, build a community, and act as a role model for privacy engineering practices and Meta values Disseminate findings through internal publications, knowledge sharing, and contributions to the broader privacy and security community

Qualifications

B.S. or M.S. in Computer Science, Cybersecurity, or related field, or equivalent experience 8+ years of hands-on engineering experience in technical privacy, data protection, or privacy-adjacent security engineering Deep expertise in at least one privacy subdomain with demonstrated ability to advance the state of the art: purpose limitation, retention and deletion, data access control, jurisdictional obligations, de-identification, consent, or ML and model privacy Experience implementing privacy obligations in production systems, in addition to reviewing or advising on them Software engineering proficiency, including code review and maintenance, with experience coding in Python, PHP, Java, C/C++, or an equivalent language Track record of architecting privacy systems, controls or programs that operate at scale and influence how others build Demonstrated ability to define and drive complex, ambiguous problem spaces with group-level impact Experience adhering to and implementing responsible, ethical AI practices (e.g., risk assessment, bias mitigation, quality and accuracy reviews) Experience improving AI model performance through expert feedback, red teaming, or evaluation design Experience with pseudonymization lifecycle, including deletion of the user-to-pseudonym mapping Demonstrated ability to integrate AI tools to optimize/redesign workflows and drive measurable impact (e.g., efficiency gains, quality improvements) Demonstrated ongoing AI skill development (e.g., prompt/context engineering, agent orchestration) and staying current with emerging AI technologies Experience implementing obligations conditioned on jurisdiction, age band, or cohort: regional carve-outs, regulatory deadlines, minor and parent-initiated request handling Contributions to the privacy or security community (original research, tools, conference presentations, publications) Experience applying differential privacy, k-anonymity, or other formal de-identification techniques in production rather than in research only Experience with privacy review or privacy incident response at scale Background in ML and model privacy: memorization, membership inference, training-data extractability, disclosure through model outputs