Research Engineer, Privacy
Research Engineer on OpenAI's Privacy team designing and prototyping privacy-preserving ML algorithms like differential privacy and federated learning at scale. Requires hands-on PETs experience, fluency in PyTorch/JAX, and a track record implementing or publishing novel privacy work.
About the job
Responsibilities
- Design and prototype privacy-preserving machine-learning algorithms (e.g., differential privacy, secure aggregation, federated learning) that can be deployed at OpenAI scale.
- Measure and strengthen model robustness against privacy attacks such as membership inference, model inversion, and data memorization leaks—balancing utility with provable guarantees.
- Develop internal libraries, evaluation suites, and documentation that make cutting-edge privacy techniques accessible to engineering and research teams.
- Lead deep-dive investigations into the privacy–performance trade-offs of large models, publishing insights that inform model-training and product-safety decisions.
- Define and codify privacy standards, threat models, and audit procedures that guide the entire ML lifecycle—from dataset curation to post-deployment monitoring.
- Collaborate across Security, Policy, Product, and Legal to translate evolving regulatory requirements into practical technical safeguards and tooling.
Requirements
- Hands-on research or production experience with PETs (Privacy-Enhancing Technologies).
- Fluent in modern deep-learning stacks (PyTorch/JAX) and comfortable turning cutting-edge papers into reliable, well-tested code.
- Enjoy stress-testing models—probing them for private data leakage—and can explain complex attack vectors to non-experts with clarity.
- Track record of publishing (or implementing) novel privacy or security work and relish bridging the gap between academia and real-world systems.
- Thrive in fast-moving, cross-disciplinary environments where you alternate between open-ended research and shipping production features under tight deadlines.
- Communicate crisply, document rigorously, and care deeply about building AI systems that respect user privacy while pushing the frontiers of capability.
Skills
Differential Privacy, Federated Learning, PyTorch, JAX, Secure Aggregation, Membership Inference, Model Inversion, Data Memorization, Privacy-Enhancing Technologies, Machine Learning, Deep Learning
Similar jobs
ML Engineering jobs
Thinking Machines Lab
San Francisco, CA
Build and operate the engineering systems that support post-training research, including reinforcement learning infrastructure, sandboxed execution, data pipelines, and agent scaffolding. The role requires strong Python and systems engineering skills, project ownership, and a relevant bachelor’s degree or equivalent experience.
Thinking Machines Lab
San Francisco, CA
Operates and improves the infrastructure powering large-scale post-training and reinforcement learning runs, partnering with researchers to debug failures, improve reliability, and automate recovery. Requires 4+ years operating distributed production systems and strong Python, Go, or C++ skills.
Thinking Machines Lab
San Francisco, CA
Research-focused engineer advancing agentic model capabilities across synthetic data, task environments, evaluations, training, and usability improvements. Requires strong Python engineering, deep learning framework experience, scalable distributed training skills, and scientific experimentation ability.
Thinking Machines Lab
San Francisco, CA
Researcher focused on scaling reinforcement learning for frontier models, with ownership spanning asynchronous RL algorithms, inference and distributed training systems, and large-scale empirical studies. Requires strong Python and deep learning experience, scalable systems debugging, and rigorous research judgment.
OpenAI
San Francisco, CA
Develop multimodal perception and authentication systems combining visual, audio, and other sensor signals for real-world AI products. The role requires machine learning expertise, practical research-to-system experience, and proficiency in Python and PyTorch with comfort in C++.