AI Researcher — Multimodal Models

Meta AI Research is seeking a researcher to advance multimodal foundation models. You will work on combining vision, audio, and language in unified model architectures, contributing to next-generation AI assistants and the Llama model family.

Responsibilities

• Design novel multimodal architectures combining vision and language
• Run large-scale pretraining and fine-tuning experiments
• Develop evaluation benchmarks for multimodal model capabilities
• Publish research at top venues (NeurIPS, ICML, CVPR, ICLR)
• Collaborate with product teams to apply research to Meta products

Requirements

• PhD in ML, Computer Vision, or NLP
• Track record of research publications
• Strong Python, PyTorch/JAX, and research engineering skills
• Experience with large-scale distributed training