Research Scientist Intern PhD, PAR Safety Evals and Red Teaming

Meta•Menlo Park, CA

4d•$7,650 - $12,667

About The Position

Meta is seeking a PhD Research Scientist Intern to join the Safety Evals and Red Teaming team within the Products and Applied Research organization. The team works on evaluating the safety of Meta’s increasingly capable AI systems. Potential research areas include agent safety, benchmarks and evaluations, synthetic data generation, automated red teaming, LLM-as-a-Judge, benchmark design, evaluation harnesses, prompt injection, and agent environments. The internship lasts 16 weeks, with various start dates throughout the year. Projects may investigate the reliability and consistency of LLM-based judges in identifying safety-policy violations; methods for evaluating prompt-injection risks in agents that use tools and untrusted information; environments and attack scenarios that reveal safety failures; the generalizability of failures found through automated red teaming; and ways to improve benchmark coverage with synthetic data while ensuring scores measure intended behavior.

Requirements

Currently has or is in the process of obtaining a Ph.D. in Computer Science, Computer Vision, Audio Processing, Artificial Intelligence, Generative AI, or a relevant technical field.
Research experience in safety, red-teaming, evaluations, machine learning, deep learning, computer vision, and/or natural language processing.
Experience with Python, C++, C, Java, or other related programming languages.
Must obtain work authorization in the country of employment at the time of hire and maintain ongoing work authorization during employment.

Nice To Haves

Experience in one or more relevant areas: model evaluation, LLM safety or alignment, agent security, prompt injection, automated red teaming, synthetic data generation, LLM-based judging, or evaluation science.
Experience designing ML experiments, analyzing complex datasets, and using quantitative approaches to investigate research questions.
Substantial research contributions through first-author publications at leading workshops or conferences such as NeurIPS, ICLR, AAAI, RecSys, KDD, IJCAI, CVPR, ECCV, ACL, NAACL, EACL, ICASSP, or similar.
Demonstrated software engineering experience through internships, employment, coding competitions, or contributions to widely used open-source projects.
Experience collaborating with teams and communicating technical results to non-technical audiences.
Intent to return to the degree program after completion of the internship or co-op.

Responsibilities

Formulate a focused research question about content safety, agent security, or the reliability of safety evaluations.
Design controlled experiments, evaluation datasets, or agent environments that test the research question and account for relevant limitations.
Implement and run experiments, including evaluation harnesses, automated attacks, synthetic data generation, or LLM-based judging methods.
Collaborate with researchers and cross-functional partners to connect research findings with safety-policy requirements and product development.
Communicate research plans, progress, and results through clear documentation and presentations, and contribute to research publications applicable to Meta’s products.