QA Analyst III - 70333-1

Request ID: 70333-1
Start/End Dates: 6/29/2026 - 6/28/2027
Tax Work Location: US - CA - Remote (105700)
Job Title: Quality Assurance - QA Analyst III

Are there any immediate disqualifiers or red flags that stand out?
No experience evaluating sales conversations or applying structured quality frameworks
Inability to articulate reasoning behind a scoring decision (written communication is critical)

Role Summary
The Ground Truth Quality Analyst is responsible for maintaining the accuracy and integrity of Meta''s AI-powered sales quality governance system. This system uses LLM-as-Judge technology to assess and evaluate Small Medium Business sales calls daily across 17 quality skills. As part of a 3-person ground truth team, you will conduct independent human reviews of AI-scored call transcripts, verify scoring accuracy, and feed findings back into the system to drive continuous improvement. Your work directly underpins the credibility of automated quality measurement at scale.

Key Responsibilities
Conduct ground truth reviews - independently evaluate a representative sample of AI-scored Mid-Market sales call transcripts against defined quality skill rubrics.
Support diagnostic deep-dives - participate in targeted investigations on calls flagged by the governance system for potential scoring issues.
Calibrate regularly - complete calibration exercises using anchor examples (clear PASS, clear FAIL, borderline) to maintain scoring consistency across the team.
Track inter-rater reliability - log review outcomes in the governance review tool and flag disagreements with AI scores for resolution.
Categorise disagreements by governance pillar - classify scoring discrepancies across Score Accuracy, Attribution Integrity, Bias & Fairness, Consistency, Coverage, and Drift to identify systemic patterns.
Feed findings into prompt refinement - document and communicate patterns in AI scoring errors to support the prompt engineering cycle and improve model performance.
Expand the ground truth dataset - contribute reviewed and validated call assessments to the growing ground truth corpus used for model benchmarking and governance reporting.

Required Qualifications
5+ years of experience in QA, auditing, or quality review - ideally in a sales, contact centre, or customer-facing environment.
Demonstrated analytical skills with a strong attention to detail and ability to apply rubrics consistently.
Experience evaluating sales conversations (calls, transcripts, or recordings) against structured quality frameworks.
Comfort working with AI-generated outputs - ability to critically assess automated scores rather than default to them.
Strong written communication skills for documenting findings and flagging patterns clearly.
Self-directed and reliable in a remote working environment with minimal supervision.

Preferred Qualifications
Experience with LLM evaluation, AI output review, or human-in-the-loop quality processes.
Experience in Quality Assurance, performance assessments etc.
Background in sales coaching, enablement, or sales methodology (e.g., MEDDIC, Challenger, SPIN).
Proficiency in multiple languages relevant to EMEA Mid-Market sales.
Experience with inter-rater reliability measurement or calibration frameworks.

Must-Have Skills
1. Quality review / QA auditing experience (5+ years) - strong attention to detail and consistency applying rubrics.
2. Sales call / transcript evaluation experience - ability to assess sales conversations against a structured quality framework.
3. Analytical + written communication skills - clearly document findings, disagreements, and patterns in AI scoring errors.

Nice-to-Have Skills
1. LLM evaluation / human-in-the-loop AI QA experience (reviewing AI outputs, LLM-as-judge, prompt evaluation).
2. Calibration + inter-rater reliability experience (agreement tracking, consistency frameworks).
3. Sales methodology / coaching background (e.g., MEDDIC, Challenger, SPIN) and/or EMEA language proficiency.