Research Scientist Intern, Core AI - Language (PhD)

The Core AI group at Reality Labs Research is seeking an outstanding Research Scientist Intern for 2026 to advance the science and engineering of large-scale LLM training, efficient LLM algorithms, small LLMs, reasoning capabilities, and pre-/post-training methods. As an intern, you will lead impactful research projects and develop reproducible training and evaluation pipelines in a top-tier industrial research environment.Your research project includes but not is limited to:- Large-scale LLM training and data curation- Efficient LLM algorithms - Small LLMs (design, training, deployment)- LLM reasoning and agentsOur internships are twelve (12) to twenty four (24) weeks long and we have flexible start dates in 2026.

Research Scientist Intern, Core AI - Language (PhD) Responsibilities:

• Plan and execute ambitious research to advance LLM pre-training and post-training (reasoning, alignment, efficiency, and compact model design).
• Build scalable pipelines for data curation, training, and evaluation
• run rigorous ablations and analyses to generate actionable insights.
• Develop reproducible experimentation and contribute high-quality code, documentation, and research communications.
• Publish and present research findings to internal and external audiences
• collaborate across research, engineering, and product teams.

• Strong programming skills in Python and experience with modern deep learning frameworks (e.g., PyTorch)
• Research and/or work experience in at least one of the following: 1) Large-scale LLM pre-training/post-training, data curation, 2) Efficient LLM algorithms, 3) Small model architecture and training, 4) LLM reasoning or post-training (instruction tuning/reinforcement learning), 5) LLM agent systems (tool-use, planning, multi-step workflows, evaluation)
• Demonstrated research impact (e.g., publications, patents, open-source contributions, awards/fellowships), as well as publications at leading conferences such as NeurIPS, ICLR, ICML, ACL, EMNLP
• Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment

• Experience with distributed training and large-scale experimentation
• Experience with efficiency techniques (e.g., quantization or distillation) and/or long-context optimizations
• Experience building systems based on machine learning, reinforcement learning and/or deep learning methods
• Demonstrated SWE or research experience via internships, competitive benchmarks, open-source tools, or widely used libraries