About this role
Amazon seeks a Machine Learning Data Associate to produce and validate annotated data used to train, evaluate, and improve machine learning models and automated customer support experiences. The role works under the guidance of subject matter experts and supports projects involving customer conversations, chat, voice, and self-service interactions.
Responsibilities
- Annotate and label data for model training and fine-tuning, including intent and dialogue labeling and multi-turn free-text annotation. Write simulated conversations for testing and training.
- Test customer service models against prompts to check intent detection and routing. Audit question-and-answer pairs across marketplaces for accuracy and policy compliance, and compare call audio with transcripts to identify mismatches.
- Review customer contacts to identify defects, assess sentiment, and find improvement opportunities. Flag personally identifiable information exposure and policy violations, and evaluate automated agents’ response templates.
- Maintain quality standards, meet throughput targets and other performance metrics, collaborate with project leads and the operations team, and participate in upskilling across project types.
Required qualifications
Fluent spoken, written, and reading English; a bachelor’s degree or equivalent; customer-facing experience and a commitment to service, or experience demonstrating strong analytical skills, attention to detail, and effective communication. Candidates must work proactively and independently, meet deadlines, and deliver assigned tasks.
Preferred qualifications
Experience with natural language or linguistic data labeling, annotation, or other data markup; multilingual speech and text data; customer insights; and managing multiple assignments or completing complex, time-sensitive work with little guidance. Experience in machine learning, data mining, information retrieval, statistics, natural language processing, or computer architecture is preferred, as is familiarity with evaluating conversational or automated customer service experiences.