Senior Model Performance Engineer

Requirements:

• MSc or PhD in Computer Science, Computer Engineering, or a related field with 8+ years of experience in machine learning, computer vision, or equivalent experience
• A passion for optimizing AI models and understanding low-level GPU processes
• Strong programming skills in Python, C++, and CUDA (Slang is a plus)
• Experience with running and optimizing distributed training on a multi-node multi-GPU system
• Strong software engineering fundamentals including source control, testing/validation, and containerization
• Excellent communication and interpersonal skills

Nice to Haves:

• Strong coding architecture skills demonstrated through contributions to large internal or open-source projects
• Hands-on experience with NCCL, MPI, or UCX communication libraries
• Experience with advanced CUDA and Slang optimization for graphics or vision applications
• History of multidisciplinary creativity and innovation with hardware and software projects in graphics or robotics
• Experience with robotic systems such as autonomous vehicles or humanoid robotics

What you'll be doing:

• Profile model training and inference to identify bottlenecks regarding efficiency, latency, and memory usage
• Optimize models with respect to GPU utilization and throughput by improving kernel efficiency
• Work with researchers to scale prototypes of models for generative video creation, segmentation, 3D reconstruction, and more
• Collaborate within a large codebase building robust models for synthetic data and world generation
• Interact closely with different research, performance, and product teams at NVIDIA
• Contribute to NVIDIA NuRec and other core NVIDIA products and libraries

Perks and Benefits:

• Highly competitive salaries
• Comprehensive benefits package
• Equal opportunity employer promoting diversity and inclusivity
• Accommodations for individuals with disabilities during the application process