Senior DL Engineer — Inference & Model Optimization

NVIDIA Corporation; Santa Clara (CA); On-site; USD 184,000 - 357,000; Full time

Benefits offered by this job

Equity

Benefits package

Competitive salary

Job summary

NVIDIA is seeking a Senior Deep Learning Software Engineer to scale automated inference and deployment for generative AI models in Santa Clara, CA. You will train and deploy state-of-the-art LLMs and diffusion models, using the Torch 2.0 stack to extract graph representations and optimize inference with advanced parallelism and kernels.

Work across teams to push performance, profiling GPU kernels and refining deployment strategies for TRT and related tools.

Qualifications

• Masters, PhD, or equivalent in CS/AI or related field.
• 8+ years of deep learning experience.
• Strong software design, debugging, and performance analysis.
• Proficiency in Python, PyTorch, and ML tools (e.g. HuggingFace).
• Strong algorithms and programming fundamentals.
• Excellent written and verbal communication; able to work independently and with teams.

Responsibilities

• Train, develop, and deploy state-of-the-art generative AI models like LLMs and diffusion models.
• Leverage Torch 2.0 ecosystem to analyze models and extract graph representations.
• Develop high-performance inference optimization techniques and tensor parallelism.
• Collaborate across teams to implement efficient kernel implementations.
• Profile GPU kernel performance and identify optimization opportunities.
• Innovate on inference performance to maintain leadership in NVIDIA's software stack.
• Architect modular, scalable software platforms with broad model support.

Skills

Deep Learning

Python

PyTorch

Algorithms

Software Design

Communication

HuggingFace

Education

Masters or PhD in CS/AI or related field

Tools

CUDA

Triton

CUTLASS

Job description

NVIDIA is seeking a Senior Deep Learning Software Engineer to scale automated inference and deployment for generative AI models in Santa Clara, CA. You will train and deploy state-of-the-art LLMs and diffusion models, using the Torch 2.0 stack to extract graph representations and optimize inference with advanced parallelism and kernels.

Work across teams to push performance, profiling GPU kernels and refining deployment strategies for TRT and related tools.