Role overview: Candidate will be working with, testing, and training AI and will be responsible for building linguistic content that enables our client’s content moderation solutions to detect and classify a wide variety of online harms with cultural context.

Responsibilities relating to Maintenance and Future Development

Building and maintaining their language of expertise in tools, using your knowledge of both the language and the cultures which speak this language to ensure that classification is accurate, and context based. Classifying words/phrases based on a defined taxonomy of online harms. Reviewing chat data to ensure accurate classification and address misclassifications as needed. Analyzing data for trends and creating classification improvement plans based on impact. Identify and add vocabulary based on client requests, global issues, current events, and other trends. Collaborating with the Client Success team to answer client questions and assist in tackling problems they may be facing within their communities. Participating in developing new and better ways of detecting content intended to circumvent content moderation. Support product testing when needed and identify and report product outages

Responsibilities relating to AI

Data Generation Using various tools to curate and build datasets for evaluating and testing models and/or prompt effectiveness Labeling or analyzing data for our data scientists to assist in training AI Red-Teaming Test for potential harms in new AI features, evaluate response, report results Work with internal teams to refine policy input for AI Contribute native language prompts to red-teaming library for future use Quality Audits Label audit data provided Report results, identify areas of weakness and recommendations for improvement