Research Scientist Intern: Efficient LLMs & Vision Models

Meta; Menlo Park, Burlingame (CA, CA); On-site; USD 55,000 - 83,000; Full time

Job summary

Meta Reality Labs is offering a Research Scientist Intern position focused on language and vision model optimization. The intern will develop efficient large language models and foundation vision models, with emphasis on on-device performance and practical deployment.

Ideal candidates are PhD students in AI fields, experienced with PyTorch/JAX and transformers, and have a track record of research publications.

Qualifications

• PhD candidate in AI or related field.
• Must have valid work authorization at time of hire.
• Hands-on experience with deep learning algorithms and transformers.
• Experience with PyTorch and/or JAX libraries.
• Experience with Python and shell scripting.
• Experience with large-scale foundation model training and on-device optimization.
• Publications at CV/NLP/ML conferences/journals.
• Experience in LLM agents/orchestration, personalization, model compression, and efficient architectures.

Responsibilities

• Develop efficient LLMs and vision foundation models.
• Optimize models for on-device inference.
• Collaborate with Reality Labs teams on research goals.
• Contribute to publications and conference submissions.

Skills

Deep learning

Transformers

Multimodal modeling

Model compression

Efficient inference

Data-centric tuning

Education

PhD in AI or related field

Tools

PyTorch

JAX

Job description

Meta Reality Labs is offering a Research Scientist Intern position focused on language and vision model optimization. The intern will develop efficient large language models and foundation vision models, with emphasis on on-device performance and practical deployment.

Ideal candidates are PhD students in AI fields, experienced with PyTorch/JAX and transformers, and have a track record of research publications.