Research Intern: Generative Recommendations & Multimodal ML

Meta; Singapore; On-site; SGD 22,000 - 39,000; Full time

Job summary

Meta is seeking highly motivated Research Interns to advance the next generation of generative recommendation systems, combining LLMs, computer vision, diffusion models, and reinforcement learning to understand content and user preferences for personalized recommendations.

As a Research Intern, you will collaborate with researchers and engineers to explore new approaches to personalized retrieval, ranking, and generation of recommendation lists.

Qualifications

• Pursuing or holding a PhD in CS, EE, or related ML field.
• Proficient in Python and PyTorch.
• Experience with LLMs and computer vision; diffusion models and RL.
• Must have work authorization to work in Singapore.
• Intent to return to degree program after internship.
• Strong research publications at top ML/vision conferences.
• Experience developing diffusion models, flow matching, or multimodal RL.
• Experience with reward modeling or offline RL.
• Research in recommender systems, personalization, or sequential modeling.
• Contributions to open-source or reproducible research artifacts.

Responsibilities

• Collaborate with researchers and engineers to explore personalized retrieval, ranking, and recommendation list generation.
• Apply diffusion models and RL techniques to recommendation problems.

Skills

Python programming

PyTorch

LLMs

Computer vision

Diffusion models

Reinforcement learning

Research publications

Education

PhD in Computer Science / Electrical Engineering or related ML field

Tools

Python

PyTorch

Job description

Meta is seeking highly motivated Research Interns to advance the next generation of generative recommendation systems, combining LLMs, computer vision, diffusion models, and reinforcement learning to understand content and user preferences for personalized recommendations.

As a Research Intern, you will collaborate with researchers and engineers to explore new approaches to personalized retrieval, ranking, and generation of recommendation lists.