OpenAI
Research Engineer, Privacy
$500K
San Francisco · Hybrid · Full-time · Posted Jul 14, 2026
About the job
Research and build privacy-preserving machine-learning techniques, tools, and standards to protect user data across OpenAI systems while balancing model utility and performance.
What you'll do
- Design and prototype privacy-preserving ML algorithms; evaluate model resilience to privacy attacks and data leakage.
- Build internal privacy libraries, evaluation suites, and documentation; investigate and publish findings on privacy–performance trade-offs.
- Define privacy standards, threat models, and audits, and collaborate across teams to turn regulatory requirements into technical safeguards.
What they ask for
- Hands-on research or production experience with privacy-enhancing technologies (PETs).
- Fluency in modern deep-learning stacks such as PyTorch or JAX, and ability to turn research papers into reliable, tested code.
- Experience probing models for private-data leakage and communicating attack vectors clearly.
- Track record of publishing or implementing novel privacy or security work.
Nice to have
- Interest in using AI tools and agents to improve research and evaluation workflows.
Pay
$0 – $500K (USD; compensation type not specified)
Tools
PyTorch, JAX
AI focus
Develop privacy-preserving ML and evaluate privacy risks in AI models, including membership inference, model inversion, and data memorization. The role may also use AI tools and agents to improve research and evaluation workflows.
Model-written summary · the original posting is authoritative