AI Implementation Quality Analyst

Granicus•Remote,

6d•$80,000 - $105,700•Remote

About The Position

The AI Quality & Evaluation Analyst is responsible for assessing the quality, correctness, completeness, and safety of AI‑generated responses across defined use cases. This role combines hands‑on human review with structured, rubric‑based evaluation incorporating automation to ensure AI systems meet documented standards before customer implementations go live. As part of the implementation team, this role is client-facing and serves as a bridge between client expectations and system behavior, translating real-world use cases, domain context, and risk tolerance into measurable evaluation criteria and actionable feedback for product and engineering teams. The role focuses on what the AI says and does, not on model training or infrastructure performance. The analyst serves as a human quality gate, ensuring outputs are accurate, appropriate for the audience, and aligned with policy and governance requirements. This role works closely with product and domain experts to translate real‑world expectations into measurable evaluation criteria and repeatable test artifacts.

Requirements

Strong analytical judgment and attention to detail
Excellent written communication skills, including the ability to explain reasoning clearly
Experience reviewing, auditing, or evaluating structured outputs such as content, decisions, or recommendations
Comfort applying detailed guidelines and rubrics consistently at scale
Familiarity with large language models and common failure modes such as hallucinations, overgeneralization, or unsafe responses
Ability to work in spreadsheets, evaluation tools, or annotation platforms
Experience with AI evaluation, data annotation, QA, trust and safety, or policy review
Exposure to human‑in‑the‑loop workflows, or benchmarking processes
Domain expertise in a regulated or high‑risk field such as government, education, healthcare, or legal services
Experience contributing to test suites, evaluation dashboards, or quality reporting
Comfort working cross‑functionally with product and engineering teams
Experience in client-facing roles such as implementation, consulting, or solution delivery in a SaaS or regulated environment
Ability to translate between business/user needs and technical system behavior

Responsibilities

Partner directly with clients during implementation to understand use cases, success criteria, and risk tolerance
Translate client requirements into evaluation frameworks, prompt strategies, and test coverage
Act as the quality liaison between client, product, and engineering to ensure alignment pre- and post-launch
Review and score AI responses using standard rubrics
Validate against sources/ground truth; flag hallucinations, omissions, and other risks
Document findings and calibrate with peers to keep scoring consistent
Build and maintain prompt banks and golden sets (expected results)
Expand coverage for edge cases, high-risk scenarios, and real user language
Track regression/drift and feed dashboards and quality reports
Triage internal and client feedback; synthesize themes across deployments
Identify systemic risks and escalate high-impact findings with clear, client-relevant context
Partner with product/engineering to validate and verify fixes, ensuring alignment with client expectations