Research Scientist Graduate- CV/NLP/Multimodality LLM, Trust and Safety - 2026 Start (PhD): TikTok

TikTok is seeking PhD graduates for a Research Scientist position starting in 2026. This role is part of the Trust and Safety algorithm team, which is responsible for developing state-of-the-art computer vision, NLP, and multimodal models to protect the platform and its users from content and behaviors that violate community guidelines.

As a graduate scientist, you will participate in developing cutting-edge content understanding models and optimizing distributed model training frameworks.

Responsibilities
Develop computer vision models or multimodality models to recognize violation content in TikTok.

Explore cutting-edge multimodal or computer vision large models (e.g., CLIP, COCA, ViT-G, ViT-22B, etc.).

Explore the application of LLMs in business scenarios, such as pre-training, zero-shot/few-shot learning, and hard case mining.

Continuously optimize the training framework to better adapt to the training of large models.

Requirements
Minimum Qualifications:

Must be pursuing a PhD degree in Computer Science or a related technical discipline.

Related research experience in at least one of the following areas: computer vision, multimodality, or LLM.

Proficient with at least one deep learning framework (e.g., PyTorch, TensorFlow).

Excellent analytical, problem-solving, communication, and collaboration skills.

Preferred Qualifications:

Published papers in top AI conferences or journals (e.g., CVPR, ICCV, NIPS, ICML, ICLR, etc.) is a plus.