Research Scientist, Reasoning & Planning

AI Breaking Wire; San Francisco, Northern (CA, KY); On-site; USD 180,000 - 240,000; Full time

Benefits offered by this job

Equity

Top-tier medical, dental, and vision

Generous parental leave and flexible工作

Job summary

OpenAI in San Francisco seeks a Research Scientist to advance reasoning, planning, and inference-time compute in frontier models. You will lead foundational and applied RL research, drive novel training paradigms, and collaborate with infra teams to run large-scale GPU clusters.

Applicants should have a PhD or equivalent in ML/CS/Statistics and a strong publication record; experience with PyTorch and distributed training on large GPUs is essential. OpenAI offers equity and excellent benefits.

Qualifications

• PhD or equivalent practical experience in ML/CS/Statistics or related field.
• Publication record at top-tier AI/ML conferences (e.g., NeurIPS, ICML, ICLR, ACL).
• Experience with PyTorch and distributed training on large GPU clusters.

Responsibilities

• Conduct foundational and applied research in reinforcement learning, self-play, and Monte Carlo tree search for large language models.
• Design novel training paradigms to improve multi-step mathematical reasoning and coding capabilities.
• Collaborate with infrastructure engineers to scale training runs across massive GPU clusters.

Skills

Reinforcement learning research

Multi-step reasoning

Collaborative research

Education

PhD or equivalent practical experience in ML/CS/Statistics

Tools

PyTorch

Distributed training on GPU clusters

Job description

About the Role

OpenAI is seeking a Research Scientist to focus on advancing reasoning, planning, and inference-time compute capabilities in our frontier models. You will be part of a multidisciplinary team pushing the boundaries of what AI models can achieve through advanced reinforcement learning and search algorithms.

Responsibilities

• Conduct foundational and applied research in reinforcement learning, self-play, and Monte Carlo tree search applied to large language models.
• Design and execute novel training paradigms to enhance multi-step mathematical reasoning and coding capabilities.
• Collaborate with infrastructure engineers to scale training runs across massive clusters of GPUs.

Requirements

• Ph.D. or equivalent practical experience in Machine Learning, Computer Science, Statistics, or related technical fields.
• Proven publication record at top-tier AI/ML conferences (NeurIPS, ICML, ICLR, ACL).
• Extensive experience with PyTorch and distributed training on large GPU clusters.

Benefits

• Highly competitive compensation including equity.
• Top-tier medical, dental, and vision benefits.
• Generous parental leave and flexible working arrangements.