⚠️ This position appears to be closed or no longer present in the source feed.
Browse current AI training jobs →
Browse current AI training jobs →
Mercor
Disclosure: This job listing contains a referral link. AITrainingJobs may receive a referral reward if you apply or sign up through it, at no extra cost to you.
About the role
- Build answer keys for a multimodal search evaluation set
- Review photos, screenshots, documents, and conversations to define what a correct answer must contain
- Write the scoring standard rather than score model answers
What you’ll do
- Identify exact names, titles, models, venues, products, and public figures
- Transcribe facts, labels, and values precisely from the source
- Read the full conversation before writing ground truth for the final turn
- Include only what the evidence supports
- Never identify private individuals or infer sensitive attributes
- Flag unclear cases instead of guessing
You’re a good fit if you
- Are highly precise with names and numbers
- Can research unfamiliar subjects and find their correct formal names
- Follow written specifications exactly
- Have experience in annotation, ground truth, search relevance, cataloguing, or fact checking
Referral disclosure: AITrainingJobs may receive a referral reward if you apply or create an account through the links on this page. This does not affect your application or cost you anything.