PhD Intern – LLM Security Research

Job Description

Applied Research Intern – AI Safety Security – Global Tech Research Program – 2027 Start (PhD) | ByteDance

The Tone:
This is a PhD internship with ByteDance, a company founded in 2012 with a mission to inspire creativity and enrich life through a suite of more than a dozen products, including TikTok, Lemon8, CapCut, and Pico. Interns actively contribute to products and research, helping to shape the organization’s future plans and emerging technologies. This role is crucial for enhancing the security of Large Language Models (LLMs) and their applications across global products, ensuring a safe and secure digital experience for users worldwide.

The TL;DR
• Role: Internship
• Location: Flexible (Remote eligible, US location for pay transparency)
• Pay: $70 hourly
• Team: Security AI team, within the Flow AI Organization
• Mission: Systematically research adversarial testing and evaluation technologies for foundation models and agents, developing a full-link defense system based on trusted privacy computing to ensure the safe and stable development of the company’s foundation model-related business.
• Tech Stack: Python, Go, Java

What You’ll Actually Do
• Research: Systematically research attack-defense detection and evaluation methods, testing benchmarks, and toolchains for multimodal foundation models and agents.
• Develop: Advance risk defense technologies and trustworthy runtime security assurance for agent-based applications.
• Innovate: Drive training and inference technologies for foundation models and samples, securing data and model assets.
• Build: Contribute to the secure construction, performance optimization, and confidential training and inference infrastructure for foundation models.

The Must-Haves
• Background: Currently pursuing a PhD in Computer Science or a related discipline, with excellent knowledge of the theory and practice of LLMs and foundation models.
• Experience: A strong publication record at leading conferences such as NeurIPS, ICML, ICLR, ACL, or EMNLP, along with demonstrated software engineering, natural language processing, or deep learning experience.
• Skills: Excellent coding ability, familiar with data structures and fundamental algorithm skills, and proficient in Python, Go, or Java.
• Bonus: Winners of competitions such as ACM/ICPC, USACO/NOI/IOI, Top Coder, or Kaggle are preferred, alongside good communication and collaboration skills, high levels of creativity, and quick problem-solving capabilities.