Large Language Model Fine-tuning Expert

• Responsible for fine-tuning large language models to improve their performance on specific tasks and to innovate based on large models.
• Responsible for guiding the team’s research direction and engineering implementation.
• Design and implement fine-tuning algorithms, adjust hyperparameters, and optimize model performance.
• Preprocess and analyze data during the fine-tuning process to ensure dataset quality and applicability.
• Evaluate and analyze fine-tuning results, adjust models to improve their performance.
• Collaborate with team members, guide other members, and ensure smooth fine-tuning process and high-quality results.
• Deeply understand the latest large language model technology and applications, and continuously improve fine-tuning skills and knowledge.

• Master’s degree or above in Computer Science, Artificial Intelligence, Mathematics, or related fields with relevant background knowledge in machine learning, deep learning, and natural language processing.
• 5+ years of work experience in NLP-related fields and positions, with 2+ years of project management experience, responsible for major projects in well-known AI companies.
• Proficiency in mainstream large language models and successful experience in tuning them.
• Experience in practical application of human-in-the-loop reinforcement learning is preferred.
• Familiar with deep learning frameworks (such as TensorFlow, PyTorch, etc.) and commonly used model evaluation and tuning, performance acceleration techniques, familiar with commonly used AI generation model frameworks, including GAN, VAE, VQGAN/Diffusion, etc.
• Proficient in data processing and analysis skills, able to handle and analyze large-scale text datasets.
• Good teamwork and communication skills, able to work closely with other team members to complete fine-tuning tasks.
• Ability to learn quickly and solve problems, continuously learning the latest technology and solving practical problems.

Please contact shandahr@shanda.com if you are interested in any of these openings, and we also welcome help from recruiting agencies.