AI Training & Evaluation Specialist (Contractor) at Outlier & Invisible Technologies
Evaluated and annotated large language model outputs to improve factual accuracy, safety, and conversational quality. Conducted RLHF-related work by producing detailed justifications to guide model behavior and ensure alignment with project rules. Assessed prompt-response pairs for logical consistency, grammar, and adherence to complex guidelines. • Reviewed LLM responses for factuality, safety, and quality. • Wrote RLHF justifications to influence model behavior. • Checked grammar and logical consistency across prompt-response pairs. • Maintained high quality scores across multiple training campaigns.