Fine-tune large language models for real-world AI applications, optimizing post-training methods such as SFT, Reward Modeling (RM), and Reinforcement Learning (RL) for both training efficiency and user experience
Research and develop automated high-quality data generation techniques, and build efficient online data flywheel pipelines
Collaborate with engineering teams and clients to explore and deploy innovative LLM applications across domains such as content creation, education, finance, and coding
Qualifications
Education & Background
Master's degree or above in Computer Science, Artificial Intelligence, Mathematics, or related fields
Experience in mathematics or programming competitions is a plus
Experience
Several years of experience in NLP or deep learning R&D
At least 1 year of hands-on experience with LLM applications
Technical Expertise
Deep understanding of the LLM technical stack, including SFT, Reward Modeling (RM), RLHF, and data synthesis
Proficiency in Python and PyTorch
Familiarity with modern model architectures such as Transformers and Mixture-of-Experts (MoE)
Strong coding skills in Python and/or C++
Preferred Qualifications
Publications in top-tier conferences such as ACL, EMNLP, or NeurIPS related to LLMs
Experience with multimodal models and familiarity with their architectures
Familiarity with Retrieval-Augmented Generation (RAG) and agent-based systems development
