Data Science Manager, Integrity

OpenAI OpenAI · AI Frontier · San Francisco, CA · Data Science

Manager for Integrity Data Science team at OpenAI, focusing on responsible AI deployment. The role involves leading a team to build measurement systems, experimentation practices, and detection/mitigation strategies against misuse, fraud, and adversarial behaviors. Key responsibilities include driving strategy across multiple integrity domains, building analytical rigor, partnering with Product & Engineering, and evolving team structure. The role emphasizes an AI-leveraged operating mode and requires strong leadership, technical grounding in DS techniques, and experience in adversarial problem spaces.

What you'd actually do

  1. Lead and scale a high-impact Integrity Data Science team—hiring, coaching, and developing DS ICs (and potentially future managers) while setting a strong technical and cultural bar.
  2. Drive strategy across multiple Integrity domains (policy enforcement, bot detection, fraud prevention, IP theft, risk measurement, abuse prevention), balancing near-term response with durable systems.
  3. Build and institutionalize analytical rigor: clear metric frameworks, experimentation standards, monitoring/alerting, and repeatable evaluation approaches for Integrity interventions.
  4. Partner deeply with Product & Engineering to shape roadmaps, prioritize the right bets, and translate ambiguous risk signals into practical product and platform decisions.
  5. Evolve team structure and operating model as the org scales—defining ownership boundaries, improving processes, and creating leverage through better tooling and AI-assisted workflows.

Skills

Required

  • Leading and scaling Data Science teams
  • Trust & safety, fraud/abuse, security, risk, or other adversarial problem spaces
  • Modern DS techniques (experimentation, causal inference, anomaly detection, risk modeling, measurement design)
  • Building durable partnerships across DS, Engineering, Product, and Operations
  • Hiring, mentoring, and developing technical talent
  • Translating threats into clear frameworks, metrics, and decisions
  • Operating in ambiguity
  • Communication with senior leadership

Nice to have

  • Deploying scaled detection solutions using LLMs, embeddings, fine-tuning, or related ML systems for abuse/fraud/risk
  • Experience with policy, content moderation, investigations, or security operations teams
  • Building or leading measurement systems that balance safety, user experience, and operational/business constraints

What the JD emphasized

  • scale the team
  • strengthen execution
  • deepen partnership
  • keep pace with fast-moving threats
  • shaping the analytical strategy
  • highly cross-functional leadership role
  • evolve team structure and operating rhythms
  • raise the bar on technical rigor
  • develop a culture of proactive, high-leverage impact
  • emergent—new misuse patterns appear as the technology and ecosystem evolves
  • strong judgment
  • comfort with ambiguity
  • build systems that scale
  • scale
  • balancing near-term response with durable systems
  • institutionalize analytical rigor
  • repeatable evaluation approaches
  • translate ambiguous risk signals
  • Evolve team structure and operating model
  • creating leverage
  • Enable cross-org outcomes
  • Communicate clearly with senior leadership
  • synthesizing complex tradeoffs
  • surfacing risk
  • driving alignment
  • Push the team toward an AI-leveraged operating mode
  • leading and scaling Data Science teams
  • trust & safety, fraud/abuse, security, risk, or other adversarial problem spaces in fast-moving environments
  • strong technical grounding
  • coach others to execute with rigor
  • track record of building durable partnerships
  • influence without authority
  • create shared accountability
  • hiring, mentoring, and developing technical talent
  • build a culture that is both high-bar and supportive
  • translate messy, evolving threats into clear frameworks, metrics, and decisions
  • keep the team focused on the highest-leverage work
  • comfortable operating in ambiguity
  • bring structure, clarity, and momentum
  • deploying scaled detection solutions using LLMs, embeddings, fine-tuning, or related ML systems for abuse/fraud/risk
  • worked closely with policy, content moderation, investigations, or security operations teams
  • design analytics that actually works end-to-end
  • built or led measurement systems that balance safety, user experience, and operational/business constraints

Other signals

  • Integrity Data Science
  • deploy powerful AI responsibly
  • protect OpenAI and our users from misuse, fraud, and evolving adversarial behaviors
  • detect, measures, and mitigates integrity risks at scale
  • emergent—new misuse patterns appear as the technology and ecosystem evolves
  • AI-leveraged operating mode