Key Responsibilities
• Produce high-quality labels for NLP, computer vision, and speech datasets, adhering to annotation guidelines compliance and quality thresholds. • Execute RLHF and prompt evaluation tasks (pairwise preference, scoring rubrics, safety/harms checks) to drive model performance improvement. • Perform NER, sentiment, intent, taxonomy/ontology mapping, and content safety labeling (PII, toxicity, bias, misinformation). • Conduct computer vision annotation: bounding boxes, polygons, keypoints, semantic/instance segmentation, and video event labeling. • Run QA evaluation: inter-annotator agreement checks, gold set validation, error taxonomies, and calibration sessions. • Maintain training data quality with sampling, spot checks, and issue triage in Jira/Asana. • Document edge cases, update label schemas, and collaborate with ML engineers on feedback cycles.



