About this role
Innodata (Nasdaq: INOD) is a global data engineering company. We believe that data and Artificial Intelligence (AI) are inextricably linked. Our mission is to enable the responsible advancement of artificial intelligence by providing the data, evaluation frameworks, and human expertise required to build AI systems that can be trusted at scale. We provide a range of transferable solutions, platforms, and services for Generative AI / AI builders and adopters. In every relationship, we honor our 36+ year legacy delivering the highest quality data and outstanding outcomes for our customers.
Role Overview We are seeking detail-oriented evaluators to conduct human quality evaluations for an enterprise AI customer support product on Instagram, WhatsApp, and Messenger. You will evaluate and benchmark AI model responses against complex evaluation rubrics using provided business knowledge bases.
What You’ll Own
- Model Evaluation: Review and score AI-generated customer interactions across Foundational, Experiential, and Operational dimensions (e.g., Action Fidelity, Faithfulness, Hallucination, Compliance, Tone, and Handoff).
- Intent & Fact Verification: Benchmark both informational (R1) and transactional (R2) customer queries against authoritative business sources (FAQs, product catalogs, SOPs) within the task UI.
- Quality Assurance: Participate in dual-review processes and daily calibration audits to ensure inter-rater agreement and establish ground-truth performance targets.
- Performance Targets: Deliver precise evaluation
You’ll Thrive in This Role If You Have
- Customer Service Background: Prior experience in customer service, call centers, retail, or handling customer communications via email, chat, or phone (highly prioritized).