Software Engineer, Coding Agent Harness Engineering - New College Grad 2026

NVIDIA AI; Santa Clara (CA); On-site; USD 120,000 - 180,000; Full time

Benefits offered by this job

Equity

Benefits

Job summary

NVIDIA AI is seeking a researcher to build and modify coding agent harnesses to study architectural decisions and analyze inference workload dynamics. You will collaborate across NVIDIA's hardware and software stack to extend co-design to the harness layer.

The role requires hands-on experience with frontier coding agents and a BS in Computer Science or related field, alongside a proven track record in evaluation frameworks and performance metrics.

Qualifications

• Candidates must have a proven track record in building coding agent harnesses and familiarity with evaluation frameworks and performance metrics.

Responsibilities

• Build and modify coding agent harnesses to research architectural decisions.
• Collaborate across NVIDIA's hardware and software stack to extend co-design to the harness layer.

Skills

Coding agent harnesses

AI inference

Agent evaluation frameworks

Inference performance metrics

Data collection

Data analysis

Co-design

Throughput

Codex

Claude Code

TTFT

Harbor

ITL

Education

BS in Computer Science or related field

Tools

Harbor

Codex

Claude Code

Job description

You will build and modify coding agent harnesses to research architectural decisions and analyze inference workload dynamics. Additionally, you will collaborate across NVIDIA's hardware and software stack to extend co-design to the harness layer.

Requirements

Candidates must have a proven track record in building coding agent harnesses and familiarity with evaluation frameworks and performance metrics. A BS in Computer Science or a related field is required, along with hands-on experience using frontier coding agents.

Key Skills

Coding agent harnesses, AI inference, Agent evaluation frameworks, Harbor, Inference performance metrics, TTFT, Throughput, ITL, Codex, Claude Code, Data collection, Data analysis, Research, AI-driven systems, Co-design

Benefits

Equity, Benefits