Senior Applied Research Scientist, Data Curation

What you'll do

• Develop efficient models and data pipelines to extract and curate multi-modal data for foundation model training.
• Build petabyte-scale pipelines for document parsing, HTML parsing, and various types of content deduplication.
• Optimize curation methodologies for multimodal data processed across hundred-node GPU clusters to improve training set quality.
• Create datasets, metrics, experiments, and validation scripts to establish standard research methodologies for customers.
• Assist ML Engineers in scaling pipelines to production using NVIDIA Inference Microservices and blueprints.
• Author research papers, blog posts, documentation, and training materials to communicate research findings to customers.
• Monitor and integrate the latest academic and industrial developments in data curation research.

What we're looking for

• Master's, Ph.D. or equivalent experience in data curation, document AI, information retrieval, or multimodal research.
• Track record of publication in leading conferences like CVPR, ICCV, ECCV, KDD, etc.
• 10+ years of experience developing multimodal systems across a range of models and platforms.
• Hands-on experience developing computer vision and document-extraction models, including layout analysis, OCR, and table/figure/formula extraction.
• Proven expertise managing distributed data frameworks like Ray, Spark, or Dask for massive, multi-node machine learning or data processing tasks.
• Excellent Python programming skills and strong hands-on experience with PyTorch or comparable modern deep learning frameworks.
• Excellent communication and interpersonal skills with the ability to share ideas through papers, blog posts, and GitHub.
• Kaggle Grandmaster status or a strong record of top-tier results in machine learning competitions (preferred).