Operations Specialist (Red Team Evaluation) – TikTok Safety Products

Role overview

This position sits within the team that evaluates global content safety systems from an adversarial standpoint. The goal is to uncover weak spots, assess how content protection performs across languages and cultures, and turn those findings into practical improvements.

What you will do

• Develop, refine, and standardize risk assessment frameworks, operating workflows, and SOPs for online signal review, external escalations, and intelligence-led investigations.
• Advance the red team risk taxonomy and forecasting approach by tracking regulatory shifts, emerging technologies, AIGC-related ecosystems, and threat patterns across platforms to surface risks early.
• Perform proactive threat hunting and detailed investigations using signals such as enforcement gaps, outside escalations, public sentiment, and abuse intelligence, then convert findings into clear risk reports.
• Design and improve AI-assisted risk assessment processes, including agent capability definitions, prompt design, human-in-the-loop structures, and close collaboration with engineering and machine learning partners on the Agentic Risk Assessment Platform.
• Create and maintain partnerships for external threat intelligence with security vendors, white-hat communities, industry peers, and intelligence providers to improve visibility into new threats, cybercrime trends, and abuse networks.
• Work closely with Product, Policy, Data Science, Trust & Safety, and Operations stakeholders to turn risk insights into scalable mitigation measures, stronger vulnerability handling, evaluation systems, and closed-loop governance.

Qualifications

• A bachelor’s degree or higher in Computer Science, Information Security, Risk Management, or a closely related discipline is required.
• Applicants should have practical background in at least one of these areas: Trust & Safety risk assessment or investigations, policy enforcement or content moderation strategy, online abuse mitigation, adversarial analysis or root-cause investigations, CTI, red teaming, blue teaming, penetration testing, or cybersecurity operations.
• Strong knowledge of what large language models and multimodal AI can and cannot do, plus hands-on exposure to prompt design, human-in-the-loop workflows, and cross-functional work with engineering and machine learning teams.
• Solid analytical and data-oriented problem-solving skills, including comfort with SQL and Python, building investigation dashboards, tracking metrics such as enforcement miss rates and pattern coverage, and performing root-cause analysis on large-scale online data.
• Chinese language proficiency is mandatory for communication with stakeholders in China.

Preferred background

• Experience evaluating multimodal and AIGC-related risks, including AI-generated images, deepfakes, synthetic audio/video, and AI-enabled policy violations, is an advantage.
• Familiarity with adversarial techniques, abuse pattern analysis, and scalable risk detection approaches is preferred.
• Prior experience in fast-moving environments focused on Trust & Safety, cybersecurity, risk management, or AI governance will be valuable.
• A strong interest in emerging online threats and changing abuse ecosystems is desirable.

About the company

The company is a major short-form mobile video platform with a mission to inspire creativity and bring joy. Its global headquarters are in Los Angeles and Singapore, with additional offices in cities including New York, London, Dublin, Paris, Berlin, Dubai, Jakarta, Seoul, and Tokyo.

Why join

The organization emphasizes creativity, curiosity, humility, resilience, and continuous iteration. It encourages employees to see challenges as opportunities, work as one team, and keep an “Always Day 1” mindset while building for a global community.

Diversity and inclusion

The workplace is committed to inclusion and values employees for their skills, experiences, and perspectives. It aims to reflect the global communities it serves and supports a diverse environment where all voices are respected.