Responsibilities
Team introduction: The Bytedance Seed team was founded in 2023 and is committed to finding new methods of general intelligence, pursuing the upper limit of intelligence, and contributing to the development of technology and society. The Seed team has long-term vision and determination in the field of AI. The team's research directions cover MLLM, GenMedia, AI for Science, robotics, etc. It has laboratories and positions in China, Singapore, the United States and other places. At present, the team has launched industry-leading general-purpose large models and cutting-edge multi-modal capabilities, supporting more than 50 application scenarios such as Doubao, Jimeng, and TRAE, and has opened them to enterprise customers through the Volcano engine. Third-party data shows that Doubao App ranks first in the Chinese market in terms of user count, and the average daily Token calls of Doubao's large models lead the industry. 1. Optimize & innovate RLHF algorithm training efficiency and model generalization ability 2. Implementation and application of Long CoT technology 3. Post-training algorithm for multi-modal large models (text, image, voice) 4. Construct high-quality, multi-domain data synthesis methods 5. Explore the application of LLM in emotional dialogue, creation and other scenarios.
Qualifications
1. Master's degree or above, majoring in computer, communications, mathematics and other related majors (extra points for those with mathematics and programming competitions) 2. Many years of NLP/deep learning R&D experience, at least 1 year of practical experience in large model application 3. In-depth understanding of LLM technology stack (such as RLHF, Prompt Engineering, LoRA fine-tuning, etc.) 4. Familiar with Python/PyTorch, CUDA optimization, and core network architectures such as Transformer and MoE have solid coding capabilities (Python/C++) 5. Publish LLM-related papers at conferences such as ACL/EMNLP/NeurIPS.
