Research Engineer, Privacy

OpenAI OpenAI · AI Frontier · San Francisco, CA · Security

Research Engineer focused on integrating privacy into AI systems, developing and deploying privacy-preserving ML algorithms, measuring robustness against privacy attacks, and defining privacy standards for the ML lifecycle.

What you'd actually do

  1. Design and prototype privacy-preserving machine-learning algorithms (e.g., differential privacy, secure aggregation, federated learning) that can be deployed at OpenAI scale.
  2. Measure and strengthen model robustness against privacy attacks such as membership inference, model inversion, and data memorization leaks—balancing utility with provable guarantees.
  3. Develop internal libraries, evaluation suites, and documentation that make cutting-edge privacy techniques accessible to engineering and research teams.
  4. Lead deep-dive investigations into the privacy–performance trade-offs of large models, publishing insights that inform model-training and product-safety decisions.
  5. Define and codify privacy standards, threat models, and audit procedures that guide the entire ML lifecycle—from dataset curation to post-deployment monitoring.

Skills

Required

  • differential privacy
  • federated learning
  • secure aggregation
  • membership inference attacks
  • model inversion attacks
  • data memorization leaks
  • PyTorch
  • JAX
  • deep-learning stacks
  • privacy-enhancing technologies (PETs)
  • ML lifecycle
  • threat modeling
  • audit procedures

Nice to have

  • publishing novel privacy or security work
  • experience bridging academia and real-world systems
  • experience in fast-moving, cross-disciplinary environments
  • experience shipping production features under tight deadlines
  • crisp communication
  • rigorous documentation
  • building AI systems that respect user privacy

What the JD emphasized

  • privacy-preserving machine-learning algorithms
  • differential privacy
  • federated learning
  • membership inference
  • model inversion
  • data memorization leaks
  • provable guarantees
  • cutting-edge privacy techniques
  • privacy–performance trade-offs
  • privacy standards
  • threat models
  • audit procedures
  • ML lifecycle
  • privacy or security work
  • privacy
  • security

Other signals

  • privacy-preserving machine-learning algorithms
  • differential privacy
  • federated learning
  • membership inference attacks
  • data memorization leaks
  • model inversion attacks