below marketmost similar roles pay hereabove market
This role pays more than
98%
of similar roles. Most pay
$136,350–$213,250
— the blue band above.
At the midpoint, this role pays about
$236k
versus about
$175k
for comparable roles.
Based on 240 similar postings.
Employer
About Nvidia
Nvidia is a leading designer of graphics processing units (GPUs) and system-on-chip units, powering gaming, professional visualization, data centers, and artificial intelligence workloads. Industry: Semiconductors & AI Computing
Nvidia currently has
1623 open roles
on FindRole.
Listed pay typically runs
$184,000–$287,500
across 1110 roles with salary data.
Software Engineer, OpenShell joins the NVIDIA OpenShell team to build runtime infrastructure for secure, scalable, production-grade AI agents. This role bridges research and engineering by identifying promising methods from academia and industry to integrate into enterprise tools and workflows. You will track advancements in tool use, planning, memory, and agent safety while building rapid prototypes, agent harnesses, and evaluation loops. Key responsibilities include red-teaming systems for reliability and security, and designing secure-by-default workflows for agents interacting with code and credentials. Required skills include experience with LLMs, multimodal generative models, evaluation frameworks, and inference optimization using tools like Triton, TensorRT-LLM, vLLM, Ray, and Kubernetes. You must possess a strong background in translating complex research into reusable products, benchmarks, and production systems at scale.
What does a Software Engineer earn in Remote?
Median $208800 from 354 postings across 50 companies.
Track cutting-edge research in tool use, planning, memory, evaluation, and agent safety.
Identify and translate research ideas into concrete product opportunities for OpenShell.
Reproduce and benchmark methods from academic papers, open-source projects, and internal research.
Build rapid prototypes including agent harnesses, evaluation loops, and self-improving workflows.
Design red-team harnesses to measure agent reliability, scalability, safety, and security.
Design secure-by-default workflows for agents interacting with tools, code, and enterprise systems.
Move technical ideas from initial prototype to production-grade product features.
What we're looking for
8+ years of professional practical experience in research engineering, software development, or a related technical field.
MS/PhD in Computer Science, Physics, or a related field or equivalent experience.
Strong background in turning complex research into reusable products, tools, demos, benchmarks, or production systems at scale.
Deep experience in LLMs, agent harnesses, multimodal generative models, evaluation frameworks, synthetic data generation, post-training, inference infrastructure, adversarial ML, or agent safety/security.
Demonstrated ability to drive independent technical investigation, run experiments, and communicate findings clearly.
Strong product sense and care for UX and AX for developer and agent ergonomics.
Experience with secure agent runtimes, tool sandboxing, capability-based security, or enterprise policy systems (preferred).
Experience with LLM inference infrastructure, model serving, or inference optimization using tools such as Triton, TensorRT-LLM, vLLM, SGLang, Ray, Kubernetes, or cloud GPU platforms (preferred).
AI Inference
GPU
TensorRT
Triton Inference Server
vLLM
SGLang
Linux
Containers
ROS 2
NVIDIA Jetson
NVIDIA Riva
NVIDIA Isaac
Quantization
Synthetic Data Generation
Performance Profiling
Python
Java
C#
C
C++
JavaScript
NodeJS
PHP
React
Drupal
Agile
REST API
Full-Stack Development
Cloud Development
.NET
Regression Testing
Automation
Client-Server Architecture
Modern C++
Python
C#
Java
Rust
Machine Learning
PyTorch
GPU Programming
Signal Processing
Computer Vision
Robotics
Web Service Architecture
FPGA
Numerical Methods
Mathematical Modelling
C++
Python
LabVIEW
National Instruments
cDAQ
cRIO
PXI
Data Acquisition
Distributed Systems
Network Communication
Unit Testing
Performance Optimization
Debugging
Data Storage