Lead Researcher, LLM & Multimodal AI

TikTok Inc.; San Jose (CA); On-site; USD 308,000 - 588,000; Full time

Benefits offered by this job

Medical, dental, and vision insurance

401(k) with company match

Paid parental leave

Short-term and long-term disability

Life insurance

Wellbeing benefits

10 paid holidays per year

10 paid sick days per year

17 days Paid Personal Time

Job summary

TikTok Inc. is seeking researchers in LLM, VLM and Omni Model domains who are experienced in single/multi-modality pretraining and applications, including evaluations, data processing and training recipes.

This role focuses on RL-based alignment, efficient training and inference, with real-world impact for TikTok’s content moderation, search and recommendations. You will lead incubation of next-generation LLM solutions, design experiments, and work with cross-functional teams to scale these

Qualifications

• Ph.D. in Computer Science, Data Science, AI or related field.
• Proficiency in Python, Rust, or C++ with deep learning frameworks (PyTorch, Deepspeed, Megatron, VLLM, etc.)

Responsibilities

• Lead incubation of next-generation, high-capacity LLM solutions for TikTok business.
• Design methods, tools, data recipes and experiments to advance state-of-the-art in large language models.
• Explore new model architecture and inference-efficient designs for LLM applications.
• Collaborate with cross-functional teams to plan and implement LLM projects.
• Extend insights from industry to academia.

Skills

Python

Rust

C++

Deep learning frameworks

Education

Ph.D in Computer Science / AI

Tools

PyTorch

Deepspeed

Megatron

VLLM

Job description

TikTok Inc. is seeking researchers in LLM, VLM and Omni Model domains who are experienced in single/multi-modality pretraining and applications, including evaluations, data processing and training recipes.

This role focuses on RL-based alignment, efficient training and inference, with real-world impact for TikTok’s content moderation, search and recommendations. You will lead incubation of next-generation LLM solutions, design experiments, and work with cross-functional teams to scale these