Senior Vision Language Model Engineer

NVIDIA•Santa Clara, CA

1d•$184,000 - $356,500

About The Position

NVIDIA is seeking a senior vision language model engineer to design and build agentic data and training workflows for autonomous vehicles, robotics, and medical applications. The role will help develop scalable, performant, easy-to-use dataset search platforms for physical AI developers, advancing dataset search and model training capabilities in NVIDIA products. The engineer will partner with researchers to develop and evaluate vision-language and vision-language-action models for video search and understanding, and build multimodal datasets and automated workflows spanning discovery, labeling, evaluation, and retraining. The role also involves exploring and productizing simulation and synthetic data, applying agentic AI workflows across the applied research lifecycle, and collaborating with research, model development, performance, and product teams. The position contributes to NVIDIA Cosmos Dataset Search and other core NVIDIA platforms and products.

Requirements

PhD with 4+ years, MS with 6+ years, or BS (or equivalent experience) with 8+ years of relevant experience in Computer Science, Computer Engineering, or a related technical field.
Strong background in modern deep learning, including transformer-based architectures, video modeling, and multimodal VLM/VLA or foundation models.
Experience training and deploying deep learning models on real-world datasets, including data preprocessing, distributed training, evaluation, debugging, and iterative improvement.
Experience with Python and at least one deep learning framework.
Current knowledge of research on image and video search in autonomous vehicles, healthcare, robotics, or related physical AI applications.
Fluency with agentic AI workflows across the applied research lifecycle, including prototyping algorithms and search pipelines, benchmarking, and integrating prototypes into production codebases.
Clear and effective communication skills and ability to work well in a dynamic, product- and research-focused team.

Nice To Haves

Strong publication record at top-tier conferences such as CVPR, NeurIPS, ICML, or ECCV.
Patents in video retrieval or a related field.
Strong coding architecture skills demonstrated through contributions to large internal or open-source projects.
Experience with robotic systems such as autonomous vehicles or humanoid robotics.

Responsibilities

Partner with researchers to develop and evaluate prototypes of vision-language and vision-language-action models for video search, video understanding, and related applications.
Design and implement agentic data workflows to automate data discovery, labeling, evaluation, and retraining.
Build, curate, and maintain high-quality multimodal datasets, including video, sensor, and language/action traces, for physical AI problems.
Explore and productize data sources including simulation and synthetic data.
Use agentic AI workflows across the applied research lifecycle.
Collaborate with research, model development, performance, and product teams.
Contribute to NVIDIA Cosmos Dataset Search and other NVIDIA platforms and products.