Applied Scientist Intern - Trust and Safety; Multimodal Foundation Model - Global Frontier Tec

Responsibilities

• Work on building machine learning models and systems to protect user content as part of the Trust & Safety team.
• Contribute to the development of multimodal moderation foundation models, focusing on training stability, routing optimization, cross‑modal alignment, and unified understanding & generation.
• Develop agentic moderation systems using reinforcement learning to enhance multi‑step decision making, tool‑call strategies, and interpretable closed‑loop reasoning.
• Conduct generalization testing across 200+ languages, adversarial detection of AIGC content, and design reward signals for few‑shot scenarios.
• Experiment with and advance large‑scale MoE architecture training, synthesis of high‑quality data, and integration of a flexible tool ecosystem.

• Currently pursuing a PhD in Computer Science, Data Science, Artificial Intelligence, or a related field.
• Strong understanding of cutting‑edge LLM research and practical experience in implementing advanced systems.
• Proficiency in Python, Rust, or C++ and experience with deep learning frameworks such as PyTorch, Megatron, or VLLM.
• Strong understanding of distributed computing frameworks and performance tuning for training, fine‑tuning, and inference.
• Excellent problem‑solving skills and creative mindset to address complex AI challenges.
• Published research papers or contributions to the LLM community are a plus.
• Experience with inference tuning, GPU acceleration, large‑scale AI networks, and PyTorch 2.0 is desirable.
• Familiarity with PEFT, RL, MoE, CoT, or Lang Chain is an advantage.

Compensation:
Hourly rate $57 – $57.

Benefits may vary depending on the nature of employment and country of work location. Interns receive day‑one access to health insurance, life insurance, wellbeing benefits, 10 paid holidays, and paid sick time (56 h if hired in first half of year, 40 h if hired in second half). Interns not working 100% remote may also be eligible for housing allowance.

Qualified applicants with arrest or conviction records will be considered in accordance with federal, state, and local laws, including the Los Angeles County Fair Chance Ordinance and the California Fair Chance Act. Our company believes that criminal history may affect the following duties: engaging with clients or colleagues, managing confidential information, and exercising sound judgment.

Tik Tok is committed to providing reasonable accommodations in our recruitment processes for candidates with disabilities, pregnancy, sincerely held religious beliefs, or other reasons protected by applicable laws. If you need assistance or a reasonable accommodation, please reach out to us at