TensorRT Performance Engineer

OpenTalent; California (MO); On-site; USD 150,000 - 210,000; Full time

Job summary

NVIDIA is seeking a Deep Learning Software Engineer, TensorRT Performance, to join its research and development team focused on optimizing the inference ecosystem. You will analyze bottlenecks, implement graph compiler algorithms, and improve tensor runtimes across datacenter GPUs and edge accelerators.

In this role, you will collaborate with the deep learning community to integrate TensorRT into OSS frameworks, contribute to Torch-TensorRT and TorchDynamo, and advance state-of-the-art

Qualifications

• BS/MS/PhD or equivalent in CS/CE/EE/AI.
• 2+ years of software development experience.
• Strong C++ and Python programming and software engineering skills.
• Experience with DL frameworks and inference libraries (TensorRT, PyTorch, JAX, TensorFlow, ONNX, vLLM, SGLang, FlashInfer).
• Experience with performance analysis and optimization.

Responsibilities

• Establish groundbreaking performance benchmarking methodologies and analysis workflows to identify performance issues and opportunities for NVIDIA’s inference ecosystem (TensorRT/TensorRT-EdgeLLM/Torch-TensorRT).
• Contribute features and code to NVIDIA/OSS inference frameworks, including TensorRT/TensorRT-EdgeLLM/Torch-TensorRT.
• Develop new model pipelines for NVIDIA’s inference ecosystem with optimized performance, covering quantization, scheduling, memory management, and distributed inference, to set the gold standard for Gen AI performance.
• Work with cross-collaborative teams inside and outside of NVIDIA across generative AI, automotive, robotics, image understanding, and speech understanding to set directions and develop innovative inference solutions.
• Scale performance of deep learning models across different architectures and types of NVIDIA accelerators.

Skills

C++

Python

Performance analysis

DL frameworks

Education

Bachelor's degree or higher in CS/CE/EE/AI

Tools

TensorRT

TensorRT-EdgeLLM

Torch-TensorRT

PyTorch

JAX

TensorFlow

ONNX

vLLM

SGLang

FlashInfer

CUDA

TileIR

CuTeDSL

cutlass

Triton

Job description

NVIDIA is seeking a Deep Learning Software Engineer, TensorRT Performance, to join its research and development team focused on optimizing the inference ecosystem. You will analyze bottlenecks, implement graph compiler algorithms, and improve tensor runtimes across datacenter GPUs and edge accelerators.

In this role, you will collaborate with the deep learning community to integrate TensorRT into OSS frameworks, contribute to Torch-TensorRT and TorchDynamo, and advance state-of-the-art