Research Scientist Intern, AI, Cyber Security, Safety — MSL Trust & Safety (PhD)
Description
Meta Superintelligence Labs (MSL) is committed to advancing AI in a safe, responsible, and scalable way. The PAR Trust & Safety team sits at the heart of this mission: we ensure the safety of billions of users across Meta's products and services by developing and deploying robust safety mitigations that unblock our most ambitious AI capabilities. We work on safety risk mitigation to directly enable launches of top-priority frontier models and the products built on them (e.g., Meta AI), and we develop solutions with real users in mind — monitoring after deployment, iterating, and evolving our strategy as new AI capabilities and risks emerge. As a Research Scientist Intern on the PAR Trust & Safety team, you will pioneer new AI security techniques and paradigms for superintelligent AI, working hands-on across the full research lifecycle. Our internships are twelve (12) to twenty-four (24) weeks long. What You'll Gain Hands-on experience with Meta's frontier AI models and infrastructure at scale Mentorship from leading researchers in AI cybersecurity Opportunity to publish at top-tier venues Impact on the safety of AI systems used by billions of people
Responsibilities
AI model’s defensive cyber capabilities: Leveraging frontier AI models (LLMs, agents) for automation. AI for Cyber: Leveraging frontier AI models and agents to improve cyber capabilities, such as vulnerability detection and patching. Evaluation: Develop systematic methodologies to probe frontier models’ cyber capabilities.
Qualifications
Currently has, or is in the process of obtaining, a PhD degree in Computer Science, Machine Learning, Statistics, or a related technical field Strong publication record in one or more of: AI safety, ML security, NLP, cybersecurity (top venues: NeurIPS, ICML, ICLR, ACL, EMNLP, IEEE S&P, USENIX Security, CCS, NDSS) Experience working with large language models (training, fine-tuning, or evaluation) Experience in cybersecurity, AI safety, trust & safety, integrity, red-teaming, or responsible AI Must obtain work authorization in the country of employment at the time of hire, and maintain ongoing work authorization during employment Experience with large language models, multimodal models, or agentic systems Experience with deep learning frameworks such as PyTorch Demonstrated ability to work independently and to drive a research project from concept to results Intent to return to a degree program after completion of the internship
Compensation: $7,650/month to $12,667/month + benefits