Tech Lead Research Scientist - Recommendation Foundation Models (Global E-commerce)
TikTok, Seattle, Washington
What you will do
- Own foundation models from research to production: define the modeling strategy and lead data, architecture, pre-training, mid-training, and post-training through deployment.
- Advance LLM-native, event-sequence-driven recommendation: model user behavior and intent to improve retrieval, ranking, and end-to-end generative recommendation.
- Build multimodal representations and semantic tokenizers that connect product, content, and behavioral signals and transfer across scenarios.
- Deliver measurable online impact: establish rigorous evaluations and experiments, balancing recommendation quality and long-term user value with latency and inference cost.
- Lead through hands-on execution: guide model and system design, align research and engineering partners, and mentor colleagues while staying close to code and experiments.