NVIDIA
Senior Software Engineer, Open Harness Engineering
$184K–$356K
US, CA, Remote +4 · Senior · Full-time · Posted Oct 21, 2026
- Location
- US, CA, Remote
US, DC, Remote
US, TX, Remote
US, NY, Remote
US, CO, Remote - Experience
- 8+ years of hands-on software-engineering experience, with demonstrated technical ownership of production systems and architecture design leadership.
- Education
- Bachelor’s or Master’s degree in Computer Science, Artificial Intelligence, Applied Math, or a related field, or equivalent experience.
About the job
Build and improve open-source AI agent harnesses, runtime capabilities, and reproducible evaluation systems to increase agent quality, reliability, cost efficiency, and speed.
What you'll do
- Develop scalable evaluation systems and methods for agent harnesses across models, frameworks, and benchmarks.
- Analyze agent traces and benchmark results to diagnose failures and validate improvements through controlled experiments.
- Ship harness and runtime enhancements, including upstream open-source contributions, tests, and documentation.
What they ask for
- Bachelor’s or Master’s degree in Computer Science, Artificial Intelligence, Applied Math, or a related field, or equivalent experience.
- Expert Python skills and working proficiency in Rust, Go, C++, or TypeScript.
- Experience building or extending LLM agents, coding agents, tool-use loops, model-provider integrations, developer tools, or evaluation infrastructure.
- Experience with production systems, distributed or sandboxed execution, reproducible experiments, and maintaining open-source or developer-facing software.
Nice to have
- Contributions or substantive work in open-source agents, agent evaluation, developer tools, or related areas.
- Experience with public agent benchmarks such as SWE-bench, Terminal-Bench, or AgentBench, or trace-analysis and observability systems.
Pay
184,000 USD - 287,500 USD (base salary, Level 4); 224,000 USD - 356,500 USD (base salary, Level 5)
Tools
Python, Rust, Go, C++, TypeScript, SWE-bench, Terminal-Bench, AgentBench
AI focus
Build and evaluate AI agents and agent harnesses, diagnosing agent failures and improving reliability, quality, cost, and latency across models.
Model-written summary · the original posting is authoritative