NVIDIA is searching for a Software Engineer Intern for the research and development of Synthetic Data Generation for Deep Learning in Computer Vision domains. We are investigating and prototyping the training of Deep Learning models with synthetic data and reduce the demand of real-world annotated data. The study is to answer how synthetic data is generated to achieve comparable accuracy on Deep Learning models trained with real-world data. Secondly, optimize the synthetic data generation at scale on NVIDIA latest GPU power. By synthesizing difficult and rare events at will, manifesting unseen configurations on demand, this effort will greatly improve Deep Learning model robustness and performance at deployment time while reducing the cost of real-world data collection.

Built on our Omniverse platform and harnessing the power of RTX ray tracing, we're revolutionizing the world of simulation, combining enterprise-grade simulation with entertainment-quality graphics. We are looking for a strong software engineer who can leverage fundamentals of computer graphics, game engine, simulation, and machine learning to develop modular software tools, pipelines, as well as assets that can advance the field of Synthetic Data Generation for Deep Learning in Computer Vision.

What you'll be doing:

• Work in an agile software development mode to research and develop tools and assets to synthetic data generate (SDG) which is measured by Deep Learning model training.
• Take a key role in researching, designing, implementing, and maintaining the production pipeline, tools, and workflow.
• Leverage existing tools and SDKs into the pipeline to enhance features
• Develop pipeline to support building scalable SDG in retail, manufacturing, transportation domains.
• Develop tools and pipelines to animate and randomize the motion and appearance of objects of interest such as household items, people, mechanical parts in indoor and outdoor scenarios
• Design and evaluate SDG strategies with respect to Deep Learning model performance.
• Collaborate closely with NVResearch team, Omniverse team on advancing the State Of The Art in SDG based Deep Learning model development.
• Documentation, patent writing, and peer-reviewed paper publication.
• Create tools, features and engine components to required coding standards.

What we need to see:

• Pursuing an MS or Ph.D. degree in Computer Science or related field.
• Track record in computer graphics, visual effect, 3D art content creation and pipelines, automated content/data build systems.
• Excellence in both C/C++ & Python software development and proven experience building complex software systems.
• Experience in CUDA programming.
• Strong mathematical and statistical background. Publications in Computer Vision domains.
• Understanding of various gaming platform limitations with regard to visual assets.
• Experience with GIS systems such as ARCGIS CityEngine, Point cloud data management and formats

Ways to stand out from the crowd:

• Background with Universal Scene Description (USD)
• Experience with one of more tools including Omniverse, Maya/3DsMax, SideFX, Unreal Engine, Blender.
• Research experience in machine learning, deep learning.

NVIDIA is widely considered to be one of the technology world’s most desirable employers. We have some of the most forward-thinking and hardworking people in the world working for us. Are you a creative and autonomous systems engineer who loves challenges? Do you have a genuine passion for advancing the state of Digital Twin Simulation across a variety of verticals? If so, we want to hear from you. Come and join our Metropolis team where you will help build our real-time, cost-effective Digital Twin computing and simulation platform driving our success in this exciting and quickly growing field.