Research Engineer / Scientist, Pretraining Safety

OpenAI is seeking a Research Engineer/Scientist for its Pretraining Safety team, focused on building safer, more capable models before deployment. This role involves developing new techniques for predicting and evaluating unsafe behavior in early-stage models, designing safer priors, and creating architectures that allow for controllable model capabilities. The team collaborates with safety teams to ensure robust, scalable, and grounded safety foundations for AI models.

Requirements

• Experience developing or scaling pretraining architectures (LLMs, diffusion models, multimodal models, etc.)
• Comfort working with training infrastructure, data pipelines, and evaluation frameworks (e.g., Python, PyTorch/JAX, Apache Beam)
• Hands-on research – designing, implementing, and iterating on experiments
• Data-driven with strong statistical reasoning and rigor in experimental design
• Value building clean, scalable research workflows