Job description

NVIDIA has been transforming computer graphics, PC gaming, and accelerated computing for more than 25 years. It’s a unique legacy of innovation that’s fueled by great technology—and amazing people. Today, we’re tapping into the unlimited potential of AI to define the next era of computing. An era in which our GPU acts as the brains of computers, robots, and self-driving cars that can understand the world. Doing what’s never been done before takes vision, innovation, and the world’s best talent. As an NVIDIAN, you’ll be immersed in a diverse, supportive environment where everyone is inspired to do their best work. Come join the team and see how you can make a lasting impact on the world. NVIDIA’s Curator team is seeking a Senior Applied Research Scientist with experience researching, developing, and deploying deep learning models at scale across a range of modalities. You’ll join a team of Applied Research Scientists, Machine Learning and MLOps Engineers working on the next generation of data curation and document extraction pipelines for foundation-model training, including capabilities that power NVIDIA Nemotron Parse, with a focus on structured content extraction, data quality and deduplication at the petabyte-scale. At NVIDIA we’re building the foundations upon which modern LLMs are trained. Our work is a critical component in the training of Nemotron LLM models which lead the field of open LLMs. Come be a part of our world-class team building the future of Curation and Retrieval. What you’ll be doing: • Working with our team of researchers to develop efficient and performant models and data pipelines that extract and curate multi-modal data (documents, image, audio and videos) used in the training of foundation models. • Building pipelines for petabyte-scale extraction and content deduplication, including document and html parsing, fuzzy and near-duplicate deduplication, semantic deduplication, and substring deduplication. • Contributing to the expansion and optimization of curation methodologies targeting petabyte-scale multimodal data run across hundred-node GPU clusters to improve the quality of foundation model training sets. • Exploring and crafting datasets, metrics, experiments, and validation scripts to develop standard methodologies for research. These methodologies will offer customers clear guidance on which models and pipelines to apply in specific contexts. • Helping ML Engineers scale pipelines to production capability through the development of NVIDIA Inference Microservices (NIMs) and blueprints which demonstrate how to deploy NIMs in a pipeline effectively. • Writing papers, blog posts, documentation and training materials that help customers understand and take advantage of our research. • Keeping up to date with the latest developments in data curation across academia and industry. What we need to see: • Candidates with a Master's, Ph.D. or equivalent experience in data curation, document AI, information retrieval or multimodal research, along with a track record of publication in leading conferences like CVPR, ICCV, ECCV, KDD, etc. • Hands-on experience developing computer vision and document-extraction models and pipelines, including layout analysis, OCR, and table, figure, or formula extraction. Kaggle Grandmaster status or a strong record of top-tier results in machine learning competitions is a strong plus. • An understanding of the state of the art in data curation research, with a focus on multimodal content extraction and deduplication. • 10+ years of experience developing multimodal systems across a range of models and platforms. Information retrieval experience is a big plus. • Proven expertise managing distributed data frameworks like Ray, Spark, or Dask, coupled with a history of deploying massive, multi-node machine learning or data processing tasks within production environments. • Knowledge of best practices in batching, streaming, and scaling of ingestion pipelines to support real-world applications. • Excellent Python programming skills and a strong hands-on experience with PyTorch or comparable modern deep learning frameworks. • An ability to share and communicate your ideas clearly through blog posts, papers, kernels, GitHub, etc. • Excellent communication and interpersonal skills are required, along with the ability to work in a dynamic, user-focused, distributed team. A history of mentoring junior engineers and interns is a plus. Location is flexible and the team is remotely situated, focusing on NA/EU time zones. We are looking for candidates in any country where NVIDIA has an office; remote work is accepted. Widely considered to be one of the technology world’s most desirable employers, NVIDIA offers highly competitive salaries and a comprehensive benefits package. As you plan your future, see what we can offer to you and your family www.nvidiabenefits.com/ Your base salary will be determined based on your location, experience, and the pay of employees in similar positions. The base salary range is 224,000 USD - 356,500 USD. You will also be eligible for equity and benefits . Applications for this job will be accepted at least until August 31, 2026. This posting is for an existing vacancy. NVIDIA uses AI tools in its recruiting processes. NVIDIA is committed to fostering an inclusive work environment and proud to be an equal opportunity employer. As we highly value diversity in our current and future employees, we do not discriminate (including in our hiring and promotion practices) on the basis of race, religion, color, national origin, gender, gender expression, sexual orientation, age, marital status, veteran status, disability status or any other characteristic protected by law.

Requirements

Must-haves

Working with our team of researchers to develop efficient and performant models and data pipelines that extract and curate multi-modal data (documents, image, audio and videos) used in the training of foundation models.
Building pipelines for petabyte-scale extraction and content deduplication, including document and html parsing, fuzzy and near-duplicate deduplication, semantic deduplication, and substring deduplication.
Contributing to the expansion and optimization of curation methodologies targeting petabyte-scale multimodal data run across hundred-node GPU clusters to improve the quality of foundation model training sets.
Helping ML Engineers scale pipelines to production capability through the development of NVIDIA Inference Microservices (NIMs) and blueprints which demonstrate how to deploy NIMs in a pipeline effectively.
Writing papers, blog posts, documentation and training materials that help customers understand and take advantage of our research.
Keeping up to date with the latest developments in data curation across academia and industry.
An understanding of the state of the art in data curation research, with a focus on multimodal content extraction and deduplication.
10+ years of experience developing multimodal systems across a range of models and platforms. Information retrieval experience is a big plus.
Proven expertise managing distributed data frameworks like Ray, Spark, or Dask, coupled with a history of deploying massive, multi-node machine learning or data processing tasks within production environments.
Knowledge of best practices in batching, streaming, and scaling of ingestion pipelines to support real-world applications.
Excellent Python programming skills and a strong hands-on experience with PyTorch or comparable modern deep learning frameworks.
An ability to share and communicate your ideas clearly through blog posts, papers, kernels, GitHub, etc.
Location is flexible and the team is remotely situated, focusing on NA/EU time zones. We are looking for candidates in any country where NVIDIA has an office; remote work is accepted.
You will also be eligible for equity and benefits .
Applications for this job will be accepted at least until August 31, 2026.
This posting is for an existing vacancy.
NVIDIA uses AI tools in its recruiting processes.

Nice-to-haves

Excellent communication and interpersonal skills are required, along with the ability to work in a dynamic, user-focused, distributed team. A history of mentoring junior engineers and interns is a plus.

Application prep

Use the role’s language and application details while tailoring your materials.

Keywords to mirror

aimachine learningllmmlpython

Platform: eightfold
Effort: medium

Benefits

Benefits detected in the listing.

Vision