LLM Research Scientist: Pretraining, Tuning & Reasoning

ByteDance; San Jose (CA); On-site; USD 244,800 - 450,000; Full time

Benefits offered by this job

Medical, dental, vision insurance

401(k) with company match

Paid parental leave

Disability coverage

Paid holidays and personal time

Job summary

ByteDance’s Seed LLM team is advancing the next generation of large language models, tackling pre-training, post-training, inference, memory, learning, and interpretability. The role focuses on exploring models, data construction, instruction tuning, and optimization to improve reasoning and related abilities.

The ideal candidate is pursuing a PhD with a 2027 graduation window and strong coding skills in C/C++ or Python, plus hands-on experience with large-scale ML systems and collaborative

Qualifications

• Pursuing a PhD with expected graduation in 2027 and able to start by end of 2027.
• Excellent coding ability, data structures and fundamental algorithms; proficient in C/C++ or Python.
• Experience with large-scale model training, deep learning, and ML algorithms.
• Excellent problem analysis and solving skills for large-scale model training and applications.
• Candidates who have led influential projects or papers in the field of large-scale models are preferred.
• Strong communication and collaboration skills to explore new technologies with the team.

Responsibilities

• Explore large-scale models and optimize associated systems.
• Perform data construction, instruction tuning, preference alignment, and model optimization.
• Improve model capabilities such as reasoning, code, mathematical reasoning, and more.
• Conduct in-depth research and exploration of future use cases.

Skills

C/C++

Python

Algorithms

Data structures

Education

PhD in computer science/related field

Job description

ByteDance’s Seed LLM team is advancing the next generation of large language models, tackling pre-training, post-training, inference, memory, learning, and interpretability. The role focuses on exploring models, data construction, instruction tuning, and optimization to improve reasoning and related abilities.

The ideal candidate is pursuing a PhD with a 2027 graduation window and strong coding skills in C/C++ or Python, plus hands-on experience with large-scale ML systems and collaborative