ByteDance
Research Scientist Graduate (Security-Flow) - 2026 Start (PhD)
San Jose, California · Full-time · Posted Oct 28, 2026
- Education
- Final-year PhD or recent PhD graduate; no specific degree field beyond the listed technical disciplines.
About the job
Research and develop security solutions for Large Language Models (LLMs), assess AI security risks, and help teams adopt secure AI practices.
What you'll do
- Design and develop LLM security solutions, including defenses against prompt injection and private information disclosure.
- Assess LLM security risks and recommend mitigations; evaluate and improve internal AI security processes.
- Collaborate with external experts and provide guidance and training on secure AI practices.
What they ask for
- Final-year PhD student or recent PhD graduate in AI, LLM, Natural Language Processing, Computer Vision, Security, or a related technical discipline.
- Excellent knowledge of LLM and machine learning theory and practice.
- Good communication and collaboration skills.
- Creativity and strong problem-solving capabilities.
Nice to have
- Strong publication record at leading conferences such as NeurIPS, ICML, ICLR, ACL, or EMNLP.
- NLP, computer vision, or deep learning experience through internships, work, or coding competitions.
Tools
Large Language Models (LLMs), machine learning, Natural Language Processing, Computer Vision
AI focus
AI safety and security research, including LLM robustness, risk assessment, and secure AI practices.
Model-written summary · the original posting is authoritative