Senior AI Agent Runtime Engineer

NVIDIA Corporation; California; Hybrid; USD 184,000 - 288,000; Full time

Job summary

NVIDIA OpenShell Research Engineer is responsible for translating cutting-edge AI research into secure, production-grade agent runtimes. You will explore agentic systems, build prototypes, and drive adoption of enterprise-grade workflows across teams.

Ideal candidates have 8+ years in research engineering or related fields, an MS/PhD in CS/Physics or equivalent, and deep experience with LLMs, agent harnesses, and inference infrastructure. Equity is part of the package.

Qualifications

• 8+ years of professional experience in research engineering, software development, or related technical field.

Responsibilities

• Track the cutting edge of agentic systems and how they evolve.
• Bridge research and product by translating ideas into concrete opportunities.
• Benchmark and reproduce methods from papers, open-source work, and NVIDIA research.
• Build rapid prototypes and agent harnesses for OpenShell workloads.
• Design red-team harnesses to measure reliability, safety, and security.
• Secure the workflow with secure-by-default enterprise practices.
• Collaborate across research, engineering, product, design, and developer-facing teams.

Skills

LLMs

Agent harnesses

Evaluation frameworks

Security awareness

Collaboration

Education

MS/PhD in CS/Physics

Tools

Triton

TensorRT-LLM

vLLM

SGLang

Ray

Kubernetes

Job description

NVIDIA OpenShell Research Engineer is responsible for translating cutting-edge AI research into secure, production-grade agent runtimes. You will explore agentic systems, build prototypes, and drive adoption of enterprise-grade workflows across teams.

Ideal candidates have 8+ years in research engineering or related fields, an MS/PhD in CS/Physics or equivalent, and deep experience with LLMs, agent harnesses, and inference infrastructure. Equity is part of the package.