Senior Multimodal Data Curation Scientist

NVIDIA; New York (NY); On-site; USD 224,000 - 357,000; Full time

Benefits offered by this job

Equity

Benefits package

Job summary

NVIDIA is seeking a Senior Applied Research Scientist to research, develop, and deploy deep learning models at scale for data curation and document extraction. You will join a team of researchers and ML engineers working on petabyte-scale pipelines across modalities to support Nemotron Parse and foundation-model training.

The role emphasizes publishing results, contributing to data quality and deduplication, and collaborating with customers to guide model and pipeline selection.

Qualifications

• Master's or PhD in data curation, document AI, information retrieval or multimodal research, with publications in top conferences.
• Hands-on experience with computer vision and document-extraction models, including layout analysis, OCR, and table/figure/formula extraction.
• Proven expertise in multimodal content extraction and data deduplication.
• 10+ years building multimodal systems across models and platforms; IR experience a plus.

Responsibilities

• Develop efficient models and data pipelines for multi-modal data extraction for foundation model training.
• Build petabyte-scale extraction/deduplication pipelines, including HTML parsing and semantic deduplication.
• Scale pipelines to production with NVIDIA Inference Microservices (NIMs) and deployment blueprints.
• Write papers, blogs, docs, and training materials to help customers use our research.

Skills

Python programming

PyTorch

Multimodal data

Research publications

Education

Master's degree or PhD in a relevant field

Tools

Ray

Spark

Dask

Job description

NVIDIA is seeking a Senior Applied Research Scientist to research, develop, and deploy deep learning models at scale for data curation and document extraction. You will join a team of researchers and ML engineers working on petabyte-scale pipelines across modalities to support Nemotron Parse and foundation-model training.

The role emphasizes publishing results, contributing to data quality and deduplication, and collaborating with customers to guide model and pipeline selection.