Join Rex.zone to produce high-accuracy training data and evaluations for AI/ML systems. You will contribute to LLM training pipelines through data labeling, RLHF preference ranking, prompt/response evaluation, and QA evaluation to improve model behavior, safety, and helpfulness.
What You Will Do
- Perform senior-level data annotation and evaluation across multiple domains while balancing speed and training data quality
- Apply detailed rubrics and maintain strong annotation guidelines compliance
- Provide clear written rationales on ambiguous cases and escalate guideline gaps
- Participate in calibration sessions and periodic gold-task checks
Core Workstreams
- RLHF & LLM evaluation: rank model outputs; assess helpfulness/harmlessness; apply policy-based judgments
- NLP labeling: named entity recognition, intent classification, text span labeling, prompt evaluation
- Computer vision annotation: bounding boxes, polygons, keypoints, segmentation masks, visual QA
- Content safety labeling: classify sensitive content and enforce safety taxonomies
- Training data QA: audits, disagreement analysis, sampling/rework, discrepancy analysis
Required Qualifications
- Mid-senior experience in data annotation or evaluation operations
- Strong reading comprehension and decision consistency under detailed rubrics
- Comfort working with NLP, LLM evaluation, and QA evaluation methods
- Ability to meet accuracy targets and throughput expectations
Compensation
$30–$50 per hour (HOURLY, base salary).
Remote, full-time role supporting distributed teams.
How To Apply
Apply through Rex.zone and be prepared to complete a short skills screening focused on training data quality, evaluation consistency, and guideline adherence.