Requirements

• Domain data extraction, performing data extraction on multimodal data such as text, tables, and images in fields such as life science, intellectual property, materials, and tele-communications;
• Build and iterate large-scale semantic search systems;
• Responsible for the research and implementation of LLM-related technologies, optimization of generative large models, and implementation in business scenarios;
• Read papers, research solutions, design experiments, and have the ability to improve existing systems independently;
• Align large models with business landing through technologies such as SFT, PEFT, DPO/PPO, etc.

Responsibilities

• Master's degree or above in mathematics and computer-related majors;
• Familiar with at least one of Python/Java/C++, with strong engineering capabilities;
• Familiar with common machine learning/deep learning algorithms, familiar with deep learning frameworks such as tensorflow, pytorch, and preferably have experience with large-scale distributed training;
• Understand the principles related to GPT, such as prompt, finetuning, transformer, and other key technologies;
• Have experience in information extraction (entity recognition, relationship extraction, event extraction), machine reading comprehension (document understanding, table understanding), information retrieval (text representation learning, hybrid retrieval), and be familiar with the principles of large models and efficient parameter fine-tuning techniques;
• Work diligently and responsibly, be good at communication, have a strong sense of teamwork, and be eager to learn.