Job Description

Contributions to the research and implementation of pre-training and alignment algorithms include ultra-large-scale multilingual pre-training technology, Mixture-of-Experts model training, Instruction Pretraining, SFT, and RLHF.
Conduct experiments to test the performance of different AI models, identifying areas for improvement and exploring new directions for enhancement.
Prospective Technical Research: Maintain a keen sensitivity to cutting-edge academic technologies (such as Vision-Language Models). Consistently publish high-quality papers at top-tier international conferences (e.g., NeurIPS, ICLR, CVPR, ICCV, and TPAMI) to drive global technical influence.
Reinforcement Learning and Reasoning Optimization: Leverage Reinforcement Learning (RL) techniques to enhance the logical reasoning capabilities of models. Design and build self-evolution loops (such as SFT-RL feedback cycles) , and optimize the end-to-end pipeline for instruction tuning (SFT) and reasoning data generation, cleaning, and structuring.

Requirement

Master's degree or higher in Computer Science, Information Technology, Programming & Systems Analysis, or other related disciplines
Experience with core techniques such as vision–language alignment or open-vocabulary recognition.
Prior full-time working experience in continual pre-training of domain-specific foundation models, such as safety, risk control, and content understanding.
Excellent problem analysis and solving skills, able to deeply solve problems in large model training and application.
Good communication and collaboration skills, able to explore new technologies with the team and promote technological progress.
Strong coding skills, data structure, and basic algorithm skills with proficiency in Python/Pytorch coding.

Pretrain Expert AI Engineer