大语言模型预训练算法工程师 - Seed Model
Core
Developing foundational large language models (Seed Model) through large-scale pretraining, continuous learning, and optimization to advance general intelligence.
Role type
Senior IC large language model pretraining algorithm engineer
Builds
Foundational LLMs supporting applications like Doubao, Jimeng, and TRAE
Domain
AI / Large Language Models / Pretraining
Deliverable
production ML models
Required skills
mathematical foundations, machine learning theory, C/C++/Python/Rust, NLP/CV algorithms, pretraining/post-training algorithms
Preferred skills
research publications in LLMs, leading influential LLM projects, Vibe coding, Agent engineering
Technologies
C, C++, Python, Rust, NLP, CV
Responsibilities
Conduct large-scale pretraining and continuous learning to overcome data and context bottlenecks; Optimize scaling mechanisms and training dynamics using mathematical and experimental methods; Improve training efficiency, convergence speed, and evaluation confidence.
