AI Research Scientist - Memory, Reasoning & Alignment Equity

Meta; New York (NY); On-site; USD 184,000 - 257,000; Full time

Job summary

Meta's FAIR RAM team seeks an AI Research Scientist to advance reasoning, memory, and alignment in long-horizon RL and agent-based learning. The role blends pure research with research-to-product work, backed by industry-scale resources and open publication.

Responsibilities include defining research agendas, running large-scale experiments, developing scalable alignment and memory techniques, and publishing in top venues.

Qualifications

• Bachelor's degree in CS/CE or related field, or equivalent practical experience.
• PhD in CS or related field.
• 2+ years of industry research in Generative AI and LLM.
• 1+ year as a formal technical lead or cross-functional impact role.
• Research in LLM post-training and/or agents.

Responsibilities

• Define and drive research agendas in reasoning, memory, alignment, and agentic learning.
• Design and run large-scale experiments to test novel algorithms (long-horizon RL, self-supervised objectives, tool-using agents).
• Develop scalable alignment methods (self-alignment, self-improvement, co-improvement).
• Invent and evaluate memory architectures for long-context reasoning.
• Publish cutting-edge research at top-tier venues and engage with the research community.
• Collaborate with other MSL teams and product-oriented pillars.

Skills

Python

PyTorch

LLM research

Leadership

Cross-functional collaboration

Education

Bachelor's degree in Computer Science or related field

PhD in Computer Science or related field

Tools

Python

PyTorch

Job description

Meta's FAIR RAM team seeks an AI Research Scientist to advance reasoning, memory, and alignment in long-horizon RL and agent-based learning. The role blends pure research with research-to-product work, backed by industry-scale resources and open publication.

Responsibilities include defining research agendas, running large-scale experiments, developing scalable alignment and memory techniques, and publishing in top venues.