AI Quality Engineer
Company: PhoenixTeam
Location: Arlington, VA (Remote)
Type: Full-time
Remote: Yes
Posted: 2026-11-06
About this role
Description
The AI Quality Engineer supports a federal customer by helping design and implement AI-enabled solutions, developing and integrating AI capabilities, and ensuring that skills, agents, and AI-generated prototypes meet quality, security, and compliance standards as they move toward production. This work supports broader modernization goals, including automating and streamlining servicing operations, reducing manual processing, improving data quality, aligning with the customer’s modernization standards, and establishing a roadmap toward the eventual retirement of legacy platforms. Success requires a bachelor’s degree (or equivalent experience), demonstrated experience in software quality assurance and test engineering, hands-on experience validating AI-generated code and prototypes, familiarity with CI/CD pipelines and quality gates, strong analytical skills, and excellent written and verbal communication.
Key Functions
- Design and implement AI-enabled solutions for customer modernization use cases, developing and integrating AI capabilities into existing systems and workflows.
- Establish and enforce a quality assurance framework — test plans, acceptance criteria, and regression suites — for validating AI-generated code, skills, and agents before they are promoted through the innovation pipeline.
- Guide the development of reusable AI skills and agents for customer modernization tasks, with particular emphasis on automated testing, security scanning, data sensitivity, and scalability.
- Collaborate with DevOps engineers to create and orchestrate long-running, multi-agent workflows, ensuring reliability, observability, and recoverability in production environments.
- Support creation of the innovation pipeline that moves AI-generated (“vibe-coded”) prototypes from sandbox concepts into secure, enterprise-grade applications.
- Support Value Engineers’ efforts by reviewing and quality-checking prototype outputs for accuracy, reliability, and ...