AI Data Annotator (Part time) at Alignerr AI (MAR 2024 to Date)
Evaluated and graded generative AI outputs for quality, accuracy, hallucinations, and overall alignment with expected behavior. Reviewed model-generated responses against strict safety, toxicity, and bias guidelines to support responsible deployment of LLMs. Authored and used high-quality prompts to simulate complex user queries and test model boundaries. • Assessed responses for factual correctness and presence of hallucinations. • Scored outputs based on alignment and quality criteria to improve model behavior. • Checked outputs for safety, toxicity, and bias compliance. • Used prompt-based scenarios to evaluate and stress-test model behavior.