Business Technology

1. Combined with vertical business scenarios, pre-train, tune and prompt optimization of large language models to ensure the availability of language models 2. Design and continuously optimize domain knowledge corpus and data annotation methods to optimize the machine learning process 3. Continue to optimize the training process and improve model training efficiency. ,1. Bachelor degree or above in computer-related majors, research direction is natural language processing, machine learning, deep learning and other related directions, or have relevant work experience in this field, and be proficient in at least one machine learning programming framework, such as tensorflow, pytorch, etc. 2. Able to abstract problems, select algorithms, optimize algorithms and continuously improve the results from complex business scenarios, and have strong self-driving power 3. Understand traditional NLP tasks such as reading comprehension, sequence annotation, text generation, and text classification. Those who have published papers in relevant top conferences such as ACL, EMNLP, KDD, SIGIR, ICML, ICLR, NeurlPS, etc. are preferred 4. Familiar with the basic principles and training methods of the industry-leading large language models (GPT series, LLaMA, ChatGLM, etc.), and have research experience in text generation and AI conversation 5. Understand the specific methods of data construction for large language model training, and be able to skillfully use machine learning algorithms to clean and structure data 6. Understand distributed training frameworks such as Deepspeed and Megatron-LM, and have certain multi-machine and multi-card distributed training and debugging experience.