Language-Based Data Annotation & Evaluation Experience on Appen and Oneforma
Worked on language based data labeling and evaluation projects on Appen & Oneforma, supporting the training and improvement of NLP and large language models (LLMs). Primary tasks included: Text classification based on intent, topic, sentiment, and quality Entity annotation (NER) for people, locations, organizations, and custom entity types Search result and response evaluation using relevance and quality metrics Text summarization and rewriting following strict guidelines Translation and localization review to ensure linguistic accuracy and cultural relevance RLHF-style rating tasks, evaluating AI-generated responses for helpfulness, correctness, and safety Consistently met quality benchmarks by strictly following annotation guidelines, handling edge cases carefully, and applying feedback from audits and reviews.