TikTok
Research Scientist Intern – TikTok Search / Generative AI (LLM) - Global Frontier Tech Recruitment Program - 2027 Start (PhD)
Share this job
Job Description
Technology
Research Scientist Intern – TikTok Search / Generative AI (LLM) - Global Frontier Tech Recruitment Program - 2027 Start (PhD)
Location
:
San Jose
Employment Type
:
Intern
Job Code
:
A29723
Responsibilities
About the team
The TikTok Search team is responsible for delivering a world-class search experience across TikTok’s global ecosystem. Our mission is to connect users with the most relevant and high-quality content through advanced technologies in search ranking, recommendation, and multimodal understanding. As a campus hire, you will collaborate with top engineers and researchers to solve challenging problems at scale, working across the full stack of search—from content understanding and retrieval to ranking and user experience optimization.
Project Overview, Challenges & Value
With the rapid evolution of pre-trained large models, TikTok Search is advancing toward a new paradigm of AI-native search. Traditional search systems face increasing limitations when handling massive-scale data, multimodal content (video, text, image), and complex user intents. This project focuses on building a next-generation generative search system powered by large models to significantly enhance search intelligence and user experience.
Key focus areas include
- Integrating large models with ranking systems to improve personalization and relevance.
- Developing end-to-end generative search models based on multimodal pretraining.
- Exploring large model–driven agent frameworks to better support complex, ambiguous, and multi-turn search scenarios.
Key challenges include
- Personalized ranking: Fully leveraging multimodal signals beyond traditional model capacity.
- Ultra-large-scale retrieval and ranking: Scaling search systems to efficiently handle hundreds of billions of candidates.
- Complex query understanding: Accurately interpreting long, ambiguous...
Keep looking