Senior High-Performance AI Engineer - CUDA Agent Systems

NVIDIA; United States; On-site; USD 152,000 - 242,000; Full time

Benefits offered by this job

Equity

Comprehensive benefits

Career growth opportunities

Job summary

NVIDIA is seeking outstanding Senior High Performance AI Engineers to build the next generation of agentic AI systems within the CUDA ecosystem. You will design, implement, and optimize agent architectures and multi-agent workflows, and develop environments for training and evaluating intelligent agents across NVIDIA's accelerated computing stack.

You will work across the stack with software, model, and hardware teams to deploy high-performance systems, leveraging CUDA and related accelerator

Qualifications

• Bachelor's degree in Computer Science, Electrical Engineering, or related field, or equivalent experience; MS or PhD preferred.
• 4+ years of relevant AI systems experience in industry or academia.
• Hands-on experience in agent systems, coding agents, reinforcement learning, or modern AI inference systems.

Responsibilities

• Design, build, and optimize agentic AI systems for the CUDA ecosystem, including agent architectures and multi-agent workflows.
• Build data, environments, verification, reward, and evaluation systems to improve model and agent capabilities.
• Co-design and optimize agentic systems across NVIDIA's software and hardware stack, from models to GPUs.

Skills

C/C++

Rust

Python

Education

Bachelor's degree in CS/EE

MS or PhD preferred

Tools

CUDA

TRT-LLM

SGLang

vLLM

Transformer Engine

Job description

NVIDIA is seeking outstanding Senior High Performance AI Engineers to build the next generation of agentic AI systems within the CUDA ecosystem. You will design, implement, and optimize agent architectures and multi-agent workflows, and develop environments for training and evaluating intelligent agents across NVIDIA's accelerated computing stack.

You will work across the stack with software, model, and hardware teams to deploy high-performance systems, leveraging CUDA and related accelerator