Research Scientist Intern, Ranking & Recommendations

About the role#

We are looking for Research Scientist Interns to join the Post-Training and Research team within Facebook Discovery. This team builds the systems that determine which content people see on Facebook, including posts, photos, videos, and Reels. You will work on ranking, recommendation, content understanding, and the post-training of AI models that power agentic experiences. Interns have the opportunity to turn open problems into research questions and work alongside research scientists and engineers to influence systems used by billions of people.

What you'll do#

• Define and lead a machine learning research project for ranking and recommendation with clear milestones.
• Design, train, and post-train deep learning models, including recommendation models, LLMs, and AI agents, using Facebook-scale data.
• Evaluate models using offline metrics and online experiments to inform the next iteration.
• Publish research results and contribute to work that applies to Meta product development.
• Collaborate with researchers and cross-functional partners to communicate research plans, progress, and results.

What you'll need#

• Current enrollment in a Ph.D. program in Computer Science, Artificial Intelligence, Machine Learning, or a related technical field.
• Experience with Python and deep learning frameworks like PyTorch.
• Familiarity with modern AI/ML techniques such as Transformers, LLMs, multimodal models, mixture-of-experts, diffusion models, RAG, and post-training methods like SFT and RL.
• Experience building systems based on machine learning or deep learning methods.
• Ability to obtain and maintain work authorization in the country of employment.
• Intent to return to your degree program after the internship.
• A track record of results through publications at venues like NeurIPS, ICML, KDD, RecSys, or similar, or through grants, fellowships, and patents.
• Research experience in areas such as LLM reasoning, generative recommendation, AI agents, or sequence and user modeling.
• Experience training large-scale models or using C++ and optimizing training and inference.
• Demonstrated software engineering experience through internships, work, coding competitions, or open-source contributions.
• Proficiency in using AI coding agents for development.

Location & details#

• Locations: Bellevue, WA; Menlo Park, CA; New York, NY.
• Work modality: On-site.
• Duration: 12 to 24 weeks with rolling start dates.
• Employment type: Full-time intern.
• Compensation: $7,650 to $12,667 per month.