AI Trainer / Data Annotation Specialist – Outlier AI (Project Aether)
As an AI Trainer and Data Annotation Specialist for Outlier AI's Project Aether, I supported large language model fine-tuning and evaluation. My work involved crafting prompts, ranking and annotating model responses, and performing comprehensive RLHF feedback using detailed rubrics. I consistently maintained high annotation quality and contributed actionable feedback to improve guidelines and model performance. • Executed prompt engineering, response ranking, and evaluation for LLM RLHF signal generation. • Identified edge cases, hallucinations, and policy violations in model outputs, supplying clear rationale for each annotation. • Conducted quality assurance reviews and maintained annotation consistency above project benchmarks. • Collaborated closely with project coordinators, influencing evaluation rubric updates through annotator feedback.