Researcher, Alignment Training

Responsibilities

• Develop synthetic data methods to teach models higher-level behavioral tendencies like intent understanding and reasoning
• Study how pre-training, mid-training, and post-training shape downstream model behavior
• Build evaluation loops connecting model behavior to training data and objectives
• Design reusable data generation and filtering pipelines for improved data quality and robustness
• Create experiments to distinguish durable learned behavior from benchmark gains or artifacts
• Collaborate across pre-training, post-training, alignment, and product teams
• Help define the research agenda for alignment training and behavioral invariance

Requirements

• Strong record of technical work in large-scale ML (pre-training, post-training, synthetic data, or evaluation)
• Ability to design experiments with subtle, noisy, or indirect signals
• Proficiency in moving between research hypothesis and engineering execution
• Strong judgment regarding research direction and signal reliability
• Clear communication skills across research, engineering, and product contexts
• Evidence-driven approach grounded in experimental results

About the Company

OpenAI is an AI research and deployment company dedicated to ensuring that general-purpose artificial intelligence benefits all of humanity.