Back to roles
DataFull-timeLocation:Shenzhen
LLM Data Engineer
Own the collection, cleaning, labeling, and versioning of multimodal data for the sports & health domain; build evaluation datasets and knowledge bases; develop automated data pipelines.
ShenzhenTeam:Data
Apply nowResponsibilities
- Collect, clean, label, augment, generate, and version-manage multimodal data for the sports & health domain, ensuring data quality and diversity.
- Design vertical evaluation datasets covering multiple tasks and difficulty levels; update evaluation data regularly to support model evaluation and iteration.
- Integrate structured/unstructured knowledge to build a professional sports & health knowledge base; maintain timeliness and accuracy to support model reasoning.
- Establish data-quality assessment standards; identify and handle noise and bias to ensure data meets training and evaluation requirements.
- Develop automated data pipelines to improve processing efficiency and support large-scale data processing.
- Collaborate with algorithm, product, and application teams to understand requirements and adjust data strategies to support business and model iteration.
- Manage data security and compliance; follow privacy regulations and handle sensitive data to ensure secure collection, storage, and use.
Requirements
- Experience in foundation-model data development; hands-on delivery experience is a plus.
- Strong programming ability; proficient in Python and the PyTorch framework.
- Excellent data insight, logical thinking, and communication; self-driven and goal-oriented.
- Master's degree or above in CS, machine learning, data science, applied mathematics, or related fields.
Send your résumé to career@speediance.com