Senior Vision Language Model Engineer
Core
Design and build agentic data and training workflows for Autonomous Vehicles, Robotics, and Medical applications, focusing on dataset search platforms and multimodal model training.
Role type
Senior IC vision language model engineer (agentic workflows)
Builds
Agentic data workflows, multimodal datasets, and core NVIDIA dataset search platforms (Cosmos)
Domain
Physical AI, Autonomous Driving, Robotics, Healthcare
Deliverable
production ML models
Required skills
transformer-based architectures, video modeling, multimodal VLM/VLA, deep learning training and deployment, distributed training, Python, deep learning frameworks, agentic AI workflows, data preprocessing, evaluation, debugging
Preferred skills
top-tier conference publications (CVPR, NeuRIPS, ICML, ECCV), patents in video retrieval, robotic systems experience, large-scale coding architecture contributions
Technologies
Python, deep learning frameworks, simulation environments, synthetic data pipelines
Responsibilities
Develop and evaluate prototypes of VLMs and VLAs for video search and understanding; Design and implement agentic data workflows to automate data discovery, labeling, and retraining; Build and curate high-quality multimodal datasets for end-to-end physical AI problems; Explore and productize new data sources including simulation and synthetic data; Collaborate with research, model development, and product teams to integrate prototypes into production codebases
