LLM Research Scientist: Scale Training & Inference

ByteDance; San Jose (CA); On-site; USD 218,000 - 480,000; Full time

Benefits offered by this job

Medical/dental/vision

401(k) match

Paid time off

Job summary

ByteDance in San Jose invites highly capable graduates to join a team building cutting-edge AI training and inference platforms. You will work with researchers to scale large language model training and reinforcement learning, focusing on GPU performance and system optimization.

This role demands a PhD candidate in a technical field, strong Python and deep learning fundamentals, and familiarity with modern distributed training frameworks.

Qualifications

• Currently pursuing a PhD in computer science, automation, electronics engineering or a related technical discipline
• Proficient in algorithms and data structures, familiar with Python
• Understand the basic principles of deep learning algorithms, be familiar with the basic architecture of neural networks and understand deep learning training frameworks such as Pytorch

Responsibilities

• Responsible for developing and optimizing LLM training & inference & Reinforcement Learning framework.
• Working closely with model researchers to scale LLM training & Reinforcement Learning to the next level.
• Responsible for GPU and CUDA Performance optimization to create an industry-leading high-performance LLM training and inference and RL engine

Skills

Python

Deep learning

Algorithms & data structures

CUDA optimization

GPU programming

Distributed systems

Education

PhD in computer science, automation, electronics engineering or related

Tools

Megatron-LM

Deepspeed

JAX

TensorRT-LLM

Pytorch

Verl

SGLang

Job description

ByteDance in San Jose invites highly capable graduates to join a team building cutting-edge AI training and inference platforms. You will work with researchers to scale large language model training and reinforcement learning, focusing on GPU performance and system optimization.

This role demands a PhD candidate in a technical field, strong Python and deep learning fundamentals, and familiarity with modern distributed training frameworks.