Senior Software Engineer - Deep Learning

NVIDIA•Us, CA

19h

About The Position

NVIDIA has been transforming computer graphics and accelerated computing for more than 25 years. In the AI era, it’s a unique legacy of innovation that’s fueled by great technology and amazing people. Our team builds state-of-the-art AI models for video streaming and broadcasting. Our models are deployed on the NVIDIA Maxine platform for real-time video communication and content creation (https://developer.nvidia.com/maxine). Our AI technologies are also offered in NVIDIA Broadcast App to enhance the video and voice in live streaming and conference calls. We are now looking for outstanding engineers to join the NVIDIA AI for media team. You will work alongside brilliant engineers on core technologies to solve ambitious computer vision and deep learning problems, especially building and optimizing real-time AI solutions that could run anywhere on cloud or premise.

Requirements

Strong experience of building and optimizing innovative AI model architectures for video use cases
Strong experience of developing efficient models with model pruning, distillation, post-quantization and quantization aware training
Experience with analyzing and fine-tuning deep learning pipeline performance
Experience with building real-time AI models for laptop and cloud use cases
Hands-on development skills using deep learning libraries and frameworks such as PyTorch/TensorFlow/ONNX, TensorRT/Triton/WinML and other neural processing SDKs
Collaboration ability to define project scope and roadmap together with the team while independently drive development effort with strong self-motivation
8+ years of relevant engineering or research background in deep learning and/or computer vision
BS, MS, or PhD in Computer Science, Electrical/Computer Engineering, or related fields (or equivalent experience)

Nice To Haves

Experience with AI inference accelerating hardware and building/optimizing models on them
Background with performance and latency analysis, profiling and tuning of AI workloads
Experience with CUDA programming, as well as a real passion for optimizing AI system performance
Experience of building platforms for computer vision such as real-time tracking of human face, gaze and body, as well as avatar animation and modeling.

Responsibilities

Develop highly efficient and low cost AI models and algorithms for computer vision and video AI
Optimize the performance, latency and power consumption of AI models on low power processors for deep learning acceleration
Deploying deep learning models and optimize the inference stack for real-time performance
Deliver the benefits of NVIDIA’s latest hardware and platform software innovations to the Deep Learning
Closely collaborate with different deep learning software and hardware teams across NVIDIA to influence roadmaps and deliver solutions