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

Join TikTok’s Algorithm team to build cutting-edge computer vision, NLP, and multimodal models that recognize and prevent trust and safety violations. You’ll develop content-understanding models, explore multimodal/computer-vision large models and LLM applications (e.g., zero-shot/few-shot, hard case mining), and continuously improve a distributed training framework. The role is designed for PhD graduates starting in 2026 with an onboarding commitment by end of year 2026.