混元大模型预训练算法研究员/专家研究员（深圳/北京/上海）
Core
Researching and developing pre-training algorithms for generative large language models, focusing on scaling laws, long-context training, and efficient knowledge embedding.
Role type
Senior IC research scientist (generative AI)
Builds
Pre-trained foundation models and AIGC capabilities
Domain
Artificial Intelligence / Large Language Models
Deliverable
production ML models
Required skills
Python, Linux, TensorFlow, PyTorch, Transformer architectures, scaling law analysis, model pre-training
Preferred skills
Experience with large-scale model pre-training, publications in top-tier machine learning conferences
Technologies
TensorFlow, PyTorch, Linux
Responsibilities
Conduct large-scale model pre-training and explore linear model structures; Research and implement efficient methods for model knowledge embedding and online learning; Analyze and predict model performance based on scaling laws under small-scale constraints.