Senior AI Data Trainer & Technical Contractor (Prolific, Remote)
Evaluated conversational AI model multi-turn outputs for logical constraints, safety thresholds, and factual clarity. Audited and debug domain-specific technical prose and structural reasoning generated by AI systems to improve dataset validation metrics. Performed targeted RLHF to minimize hallucinations in text generation models.• Assessed reasoning consistency against predefined rules and safety criteria• Rated and flagged output quality issues related to factuality and constraint adherence• Supported iterative dataset improvements via validation and error analysis• Applied RLHF-oriented feedback loops to refine model behavior