Senior Software Quality Engineer, Agentic Harness

Adobe•San Jose, CA

1d•$113,400 - $221,750

About The Position

As a Quality Engineer on the Agentic Harness team, you’ll help the harness behave the way creators expect. You’ll verify that the agent’s behavior, tool use, and outputs meet Adobe’s quality and reliability bar before they ship. You’ll work closely with engineers and data scientists to catch regressions early and harden test coverage. You’ll help the team ship quickly without breaking what already works. You’ll bring a quality perspective to a fast-moving, agent-based system. You’ll write and maintain automated test suites and design test plans for new agent capabilities. You’ll dig into failures to find root cause. This is a hands-on role where you’ll grow your skills evaluating LLM-based and agentic systems. You’ll join a team building quality practices for this space from the ground up.

Requirements

Bachelor’s degree in Computer Science or a related field, or equivalent experience.
Several years of experience in quality engineering, test automation, or QA, ideally on complex, evolving software.
Hands-on experience writing automated tests and test frameworks.
Solid grasp of test design fundamentals, including coverage, edge cases, and regression risk.
Comfort digging into bugs and failures across a stack to identify root cause, not just report symptoms.
Familiarity with testing LLM-based or agentic systems and how their behavior differs from traditional software.
Comfort operating in a fast-moving, ambiguous environment where the technology and user expectations are evolving quickly.

Responsibilities

Design, write, and maintain automated test suites and test plans covering the agentic harness’s behavior, tool use, and outputs.
Implement manual and exploratory testing on new agent capabilities, identifying edge cases and failure modes before they reach production.
Investigate test failures and reported bugs to isolate root cause, working with engineers to reproduce, document, and resolve issues.
Partner with the Agent Evals team to close coverage gaps and improve the reliability of quality signals.