大模型算法工程师（J96030）
Core
Optimize large language models for search and other Baidu business operations, focusing on MoE sparsification, pre-training task design, post-training optimization, and reinforcement learning for reasoning.
Role type
Senior IC large model algorithm engineer (LLM optimization)
Builds
Production large language models for search and enterprise applications
Domain
Search, Large Language Models, Deep Learning
Deliverable
production ML models
Required skills
Large language model principles, NLP, deep learning, algorithm optimization, Python, C++, PyTorch, MoE sparsification strategies, pre-training task design, post-training optimization, reinforcement learning, Megatron framework
Preferred skills
Search domain large model application experience
Responsibilities
Design and implement MoE sparsification strategies and pre-training tasks; Optimize inference efficiency and reduce deployment costs via MoE routing and attention sparsification; Conduct research and publish at top-tier conferences (ACL, NeurIPS, ICLR); Collaborate with data scientists and architecture engineers to deploy technical solutions.